Inertia-considered key line identification method for large-scale new energy direct current delivery system

Abstract

The invention discloses an inertia-considered key line identification method for a large-scale new energy direct-current delivery system. An inertia risk regulation factor is introduced by using a node inertia and system stability support relationship to establish a system outage probability model considering node inertia spatial-temporal distribution; on the basis of the model, the uniformity and degree of system outage probability distribution after the target line is shut down are analyzed, and a system outage probability distribution entropy index, a system outage probability change entropy index and an improved power grid topology entropy index three-dimensional identification index system are constructed in combination with a global topology structure of the system; and finally, generating a three-dimensional identification index result by reading the initial topological structure and the operation parameters of the power grid, determining the weight of each index by adopting an entropy weight method, and integrating the three-dimensional identification indexes by utilizing an improved double-layer grey key line comprehensive evaluation method, thereby carrying out identification and risk sorting on the key lines. According to the method, the problem that cascading failure risks are likely to occur in a large-scale new energy direct current delivery system in a low-inertia scene is solved.

Description

Technical Field

[0001] The present invention belongs to the field of identification of key lines in power systems, and relates to a method for identifying key lines in a large-scale new energy DC external power transmission system considering inertia. Background Art

[0002] With the proposal of the goals of "carbon peak" and "carbon neutrality" in China and the construction of a new power system with new energy as the main body, more attention has been paid to the development and utilization of new energy in desert, Gobi and hilly areas and the role of UHV DC in the external power transmission and consumption of new energy. However, for the Qishao UHV DC that transmits the power of the Jiuquan 10-million-kilowatt large-scale wind power base, the problems of grid safety and stability are prominent, and the DC has been unable to fully transmit power for a long time, seriously restricting its role in external power transmission and consumption of new energy. It is urgent to deeply examine the restrictive factors affecting the improvement of DC capacity, explore the chain fault propagation mechanism after the disturbance of the new energy external power transmission system, identify the key fault propagation modes, and enhance the ability of various resources to cooperate to ensure the large-power transmission of DC.

[0003] Due to the obvious reverse distribution characteristics of solar and wind energy resources and load centers in China, the State Grid has built multiple UHV DC power transmission external channels to send large-scale clean energy to load centers. In a high-proportion new energy DC power transmission system, the power transmission interruption caused by a DC line fault will cause a transient overvoltage phenomenon in the AC network around the rectifier station. This overvoltage may trigger the cascading tripping of adjacent new energy generating units, thereby threatening the safe and stable operation of the power system.

[0004] The research methods for identifying critical lines in current power system cascading failures are mainly divided into two categories: structural characteristic analysis and operating state identification. In terms of structural characteristic analysis methods, scholars have proposed network characteristic indicators such as node degree and electrical betweenness to identify critical components by constructing a power grid topology model. Although such methods have the advantage of computational efficiency, due to their abstract processing, they lack physical characteristics and have the defect of insufficient quantization accuracy. In terms of operating state identification, based on physical process modeling tools such as the OPA model and power flow entropy theory, weak links are identified by simulating the fault evolution path. The power flow entropy index can be used to identify critical lines in the power grid, but it should be noted that the criticality of the line will be dynamically adjusted with the change of the load rate. Research shows that when the line power flow is within the normal range, the fluctuation of the load rate in a specific interval will not affect its outage probability. This means that when evaluating the criticality of a line from the perspective of fault probability, its criticality characteristics are relatively stable. It should be noted that in the process of identifying critical lines based on the load rate, if the line power flow exceeds its transmission limit power range, the change of the load rate should no longer be used as the basis for criticality determination. This is because to prevent the line from overheating and causing a fusing accident, the system will take protective measures to cut off the line, so the change of the load rate in this state has lost its evaluation significance. Some scholars have combined the line outage probability model, but they have not considered the problem of uneven distribution of the overall system outage probability of the power grid, and most outage scenario studies are applicable to traditional power grids with strong support. In the context of a large number of new energy sources being connected to the power system currently, the characteristic of uneven distribution of the inertia level of the system is more obvious.

[0005] Existing research mostly focuses on the optimization of single-dimensional indicators and lacks a comprehensive quantification mechanism for "dynamic-static" multi-dimensional risks. Summary of the Invention

[0006] The present invention proposes a method for identifying critical lines in a large-scale new energy DC external power transmission system considering inertia, aiming to solve the problem of the risk of cascading failures in the low-inertia scenario of a large-scale new energy DC external power transmission system.

[0007] The technical solution adopted by the present invention is as follows:

[0008] A key line identification method for large-scale new energy DC external power transmission systems considering inertia. This identification method uses the relationship between node inertia and system stability support to introduce an inertia risk adjustment factor to establish a system outage probability model considering the spatio-temporal distribution of node inertia. Secondly, based on this model, analyze the uniformity and degree of the system outage probability distribution after the outage of the target line, and combine the global topological structure of the system to construct a three-dimensional identification index system including the system outage probability distribution entropy index, the system outage probability change entropy index, and the improved power grid topological entropy index. Finally, generate the three-dimensional identification index results by reading the initial topological structure and operation parameters of the power grid, use the entropy weight method to determine the weights of each index, and use the improved two-level grey key line comprehensive evaluation method to synthesize the three-dimensional identification index, so as to identify the key lines and conduct risk ranking.

[0009] The establishment of the system outage probability model considering the spatio-temporal distribution of node inertia is as follows: First, establish a line outage probability model by considering the dynamic operation characteristics of the system, including line power flow, bus voltage, system frequency, and the influence of potential hidden faults. The expression is as follows:

[0010]

[0011] In the formula: p u is the line outage probability; P L reflects the hidden fault probability of the relay protection system; P H describes the line overload risk; among the model parameters, L i represents the real-time power flow of line i; L ie is the normal operation limit of the line; L imax represents the maximum transmission power of the line;

[0012] Introduce the concept of system inertia into the above outage probability model. The expression of the total system inertia constant is:

[0013]

[0014] In the formula: G S is the total system inertia constant; m is the number of generators in the DC external power transmission system; G j is the inertia constant of generator j;

[0015] There is a specific correlation between the inertia constant of each node and the disturbance amplitude as:

[0016]

[0017] In the formula: G i is the inertia constant corresponding to system node i during the disturbance; f z is the system node frequency; Y i,j is the voltage correlation matrix between node i and generator j; Bi,j is the internal potential of generator j and the susceptance of fault node i;

[0018] When a fault risk occurs in the DC external power transmission system, it causes a break in the external power balance and then power changes. At this time, the generator needs to spontaneously release kinetic energy to provide active support for the system. The specific support expression is:

[0019]

[0020] In the formula: P h is the output power of active support; f(t) is the synchronous generator frequency;

[0021] If the frequency deviation value is larger, the support power output by the generator is higher. The relationship between the support power and the frequency change amount is:

[0022]

[0023] In the formula: G av is the generator inertia response constant; f n is the rated frequency of the generator; f e is the extreme frequency point of the generator;

[0024]

[0025] In the formula: A i is the reduction factor in the node i matrix; G min , G max are respectively the minimum and maximum values of the generator inertia constant;

[0026] The inertia parameter plays an important supporting role in the system operation. As a risk adjustment factor, it can significantly reduce the comprehensive outage probability of the system; this supporting ability is

[0027] reflected by the negative correlation between inertia and system risk: p = χp u (9)

[0028] In the formula: p is the comprehensive outage model combined with the inertia constant; p u is the line outage probability model; χ is the reduction factor matrix, and its expression is:

[0029]

[0030] The system outage probability distribution entropy index is used to characterize the relationship between line outage and power grid operation status; its setting is as follows:

[0031] Let K 1 , K 2 …K nis a sequence of constants; when a fault occurs in the target line, the outage probability of the \(i\)-th line is \(P\). i , for a given sequence of constants \(K = \{K\) 1 , K 2 … K n}, define \(I\) k as the number of lines that satisfy \(P\) i ∈ (\(K\) k , \(K\) k+1 ); then the probability \(T\) k , \(K\) k+1 ) that the line outage probability in the system falls within the interval (\(K\) k can be expressed as:

[0032]

[0033] Let the outage probability distribution vector be \(T = [T\) 1 , \(T\) 2 , … \(T\) n-1 ; if there are \(S\) k , \(K\) k+1 ) and \(P\) k denotes the outage probability of the \(i\)-th line when \(P\) i,k ∈ (\(K\) i , \(K\) k , \(K\) k+1 ), then the average outage probability \(W\) k of the lines in this interval can be expressed as:

[0034]

[0035] In the formula: \(P\) i,k is the outage probability of the \(i\)-th line that satisfies \(P\) i ∈ (\(K\) k , \(K\) k+1 );

[0036] The entropy value \(G\) x of the system outage probability distribution is:

[0037]

[0038] \(G\) x characterizes the dispersion degree of the system outage probability distribution after the target line fails; when the value of \(G\) x is relatively large, it indicates that the system is more likely to trigger cascading failures under perturbations.

[0039] The system outage probability change entropy index is used to quantify the dispersion degree of the system outage probability distribution after the target line fails, and its setting is as follows:

[0040] Assume that after line \(i\) fails, the outage probability of other line \(k\) is denoted as By comparing the initial outage probability of the system in the steady state with the new steady-state probability of the fault the change in the outage probability can be obtained and its expression is as follows:

[0041]

[0042] Then the total change in the system outage probability is:

[0043]

[0044] In the formula: L represents the total number of lines in the power grid topology;

[0045] Define the proportion η k,i of the change in the outage probability of line k as:

[0046]

[0047] In the formula: ΔP t,i represents the change in the total system outage probability; thus, the outage probability change vector η = [η 1,i , η 2,i , …, η L,i can be obtained. Furthermore, define the outage probability change entropy G c as:

[0048]

[0049] When the value of G c is small, it indicates that the impact of the fault is concentrated on a few lines, and the system is more likely to experience cascading failures.

[0050] The improved power grid topology entropy index is set as follows:

[0051] Define the degree of node n as:

[0052]

[0053] where the node connection status quantity v a,b satisfies:

[0054]

[0055] In the formula: N is the total number of system nodes;

[0056] Given the constant sequence Q = {Q 1 , Q 2 , …, Q m}, define M k as the degree R n ∈(Q k , Q k+1The number of nodes in (), then the node degree distribution probability S k is:

[0057]

[0058] From this, the node degree distribution vector S = [S 1 , S 2 , …, S m-1 can be obtained, and then the improved entropy G f of line i is defined as:

[0059]

[0060] Among them, the topological connection degree H i is calculated as follows:

[0061]

[0062] In the formula: T k is the average degree of nodes with degree R n ∈(Q k , Q k+1 ), K i is the sum of the degrees of the two end nodes of line i; G f reflects the robust characteristics of the power grid topological structure, and the larger its value, the higher the importance of the line in the power grid.

[0063] The structure of the improved two-level grey key line comprehensive evaluation method includes a criterion layer and an index layer. The system outage probability distribution entropy G x , the outage probability change entropy G c , and the improved power grid topological entropy G f are used as three indicators to evaluate the key lines in two layers;

[0064] In the criterion layer, two core evaluation criteria are established to reflect the overall stability and structural robustness of the power grid during operation. These two criteria are respectively:

[0065] 1) Operational stability C 1 layer: Evaluate the stability of the power grid under normal and disturbed states. This criterion focuses on the response ability and recovery ability of the system in the face of emergencies such as faults and load changes, ensuring that the power system can maintain continuous and safe operation;

[0066] 2) Structural robustness C 2 layer: Evaluate whether the topological structure of the power grid can effectively resist potential faults and interferences; this criterion considers the rationality of the power grid configuration, the connection between nodes and its support for system stability, and avoids wide-ranging impacts caused by single-point failures.

[0067] Using the double - level grey critical path comprehensive evaluation method, its algorithm first proposes a sliding window normalization method to process time - varying data:

[0068]

[0069] In the formula: τ = 24h is the time - window length, effectively capturing the intraday fluctuation characteristics.

[0070] Fusing the improved entropy weight method and dynamic AHP:

[0071]

[0072] In the formula: w j (t) represents the combined weight of the j - th index at time t; γ is the combined adjustment coefficient (0 ≤ γ ≤ 1) used to balance the ratio of the objective entropy weight and the subjective AHP weight;

[0073] Based on the objective weight calculated by the improved sliding entropy weight method, it reflects the information dispersion degree of the index data within the time - window [t - τ, t]:

[0074]

[0075]

[0076] In the formula: where E j (t) is the sliding information entropy, calculated by normalizing the data ; p ij (t) represents the contribution degree of the i - th sample in the j - th index; t ∈ + reflects the time characteristics of dynamic evaluation; is the dynamic weight calculated by the entropy weight method; N is the total number of samples within the sliding time - window. For every 10% increase in the new - energy proportion, N decreases by 2; Using the optimal values of each index to form the reference sequence X 0 (t), calculating the correlation coefficient γ ij (t) and the correlation degree R i (t).

[0077] When using dynamic grey relational analysis, first construct a time - varying reference sequence:

[0078] X 0 (t)=[minG x (t),minG c (t),maxG f (t)] (27)

[0079] Improve the calculation of the correlation coefficient:

[0080]

[0081] Among them, the discrimination coefficient ρ(t) is adaptively adjusted:

[0082]

[0083] A two - layer correlation degree synthesis model is established:

[0084]

[0085] In the formula: Δ ij (t) is the absolute difference between the i - th object and the reference sequence in the j - th index; Δ min All Δ ij (t) is the minimum value; Δ max All Δ ij (t) is the maximum value; Var(X(t)) is the coefficient of variation (standard deviation / mean) of the index matrix, w c (t) is the weight of the criterion layer (c = 1 for operation stability, c = 2 for structural robustness); J c The index set corresponding to the criterion layer c; among them, ρ = 0.5 is the discrimination coefficient, and the larger the R i (t) value, the higher the criticality of the line.

[0086] In view of the problem that traditional methods are difficult to accurately evaluate the criticality of lines in a system with a high proportion of new energy and significantly reduced system inertia in the sandy, gobi and desert areas, the present invention proposes a method for identifying critical lines in a large - scale new - energy DC external - power - transmission system considering inertia. This technology creatively integrates the node inertia characteristics into the critical - line evaluation model, combines the spatio - temporal distribution characteristics of system inertia in the scenario of high - penetration new energy, establishes a node inertia influence factor from the dynamic operation dimension, and constructs a fault - probability entropy model by comprehensively considering key parameters such as voltage level and component load rate, realizing the accurate quantification of the vulnerability of system operation.

[0087] Secondly, an innovative improvement scheme is proposed in the dimension of topological - structure analysis. By introducing an improved network - architecture entropy index, breaking through the single - ness defect of the traditional structure entropy, innovatively integrating topological - attribute parameters such as line betweenness and connection tightness, a new structure - entropy model that can characterize the vulnerability of the grid structure is constructed, thereby reflecting the topological importance of the line in the global network and forming a criticality evaluation benchmark in the structure dimension.

[0088] Finally, based on the "dynamic operation - static structure" two - dimension collaborative mechanism, a comprehensive identification algorithm is developed. Combining the grey - system theory to deal with the problem of incomplete information, using the entropy - weight method to dynamically optimize the index weights, ultimately realizing the identification of critical lines with multi - dimension and high robustness, and effectively solving the problem of identifying critical lines in scenarios such as the system inertia decline and enhanced volatility caused by high - proportion new - energy access, and the missing or ambiguity of index data. Brief Description of the Drawings

[0089] Figure 1 is the line outage probability model of the present invention;

[0090] Figure 2 is the structure diagram of the improved two - level grey critical line comprehensive evaluation method of the present invention;

[0091] Figure 3 is the algorithm flow chart of the improved two - level grey critical line comprehensive evaluation method of the present invention;

[0092] Figure 4 is the flow chart of the critical line identification of the present invention. Specific implementation manner

[0093] The present invention will be further described below in conjunction with the accompanying drawings and specific processes.

[0094] A method for identifying critical lines in a large - scale new - energy DC external - power - transmission system considering inertia introduces an inertia risk adjustment factor by using the relationship between node inertia and system stability support to establish a system outage probability model considering the spatio - temporal distribution of node inertia; secondly, on the basis of this model, analyze the uniformity and degree of the system outage probability distribution after the outage of the target line, and combine the global topological structure of the system to construct a three - dimensional identification index system of "system outage probability distribution entropy", "system outage probability change entropy" and "improved power - grid topological entropy"; finally, generate the results of the three - dimensional identification indexes by reading the initial topological structure and operation parameters of the power grid, determine the weights of each index by using the entropy - weight method, and comprehensively analyze the three - dimensional identification indexes by using the grey correlation analysis method, so as to identify critical lines and conduct risk ranking. The specific identification process is as Figure 4 shown.

[0095] Existing methods mainly target traditional power systems dominated by synchronous generators. In systems with a high proportion of new energy and significantly reduced system inertia, traditional methods are difficult to accurately evaluate the criticality of lines, and their research perspectives are often limited to the local influence reflected by the "degree" attribute of lines, without fully considering global characteristics. Therefore, the present invention comprehensively considers the spatio - temporal distribution of new - energy output fluctuations and node inertia, accurately quantifies the system operation risk in low - inertia scenarios, and the constructed multi - dimensional identification index system can comprehensively reflect the topological importance, power - flow distribution characteristics of lines in the global network, and their impact on system stability. Specifically as follows:

[0096] 1. Inertia node outage probability model

[0097] As Figure 1 shown, the line outage probability model adopted by the present invention has made important improvements on the basis of traditional methods. Different from conventional methods that rely on historical statistical data or steady - state assumptions, this model fully considers the system dynamic operation characteristics (including line power flow, bus voltage and system frequency) and the influence of potential hidden faults.

[0098] The specific expression of the line outage probability is as follows:

[0099]

[0100] In the formula: p u is the line outage probability; P L reflects the hidden fault probability of the relay protection system; P H describes the line overload risk; among the model parameters, L i represents the real-time power flow of line i; L ie is the normal operation limit of the line; L imax represents the maximum transmission power of the line;

[0101] Since the node inertia is positively correlated with the anti-disturbance ability of the system, in the face of the continuous increase in the proportion of renewable energy in the large-scale new energy DC external transmission system, the influence of the system inertia characteristics on the safe operation of the power grid has become increasingly prominent, making the system inertia a key indicator in the field of power grid risk assessment. Therefore, this indicator is introduced into the outage probability model. Load disturbances will cause power imbalance in the system. At this time, the inertial characteristics of the system nodes will significantly affect the frequency response characteristics: nodes with larger inertia constants show a more gentle rate of frequency change, which is mainly due to the mechanical kinetic energy stored in rotating equipment being able to effectively buffer power fluctuations. Specifically, the rotor of a synchronous generator can automatically compensate for the power difference in the system by virtue of the stored rotational kinetic energy, thereby maintaining the dynamic balance of the system. The expression of the total system inertia constant is:

[0102]

[0103] In the formula: G S is the total system inertia constant; m is the number of generators in the DC external transmission system; G j is the inertia constant of generator j.

[0104] There is a specific correlation between the inertia constant of each node and the disturbance amplitude as:

[0105]

[0106]

[0107] In the formula: G i is the inertia constant corresponding to system node i during the disturbance; f z is the system node frequency; Y i,j is the voltage correlation matrix between node i and generator j; B i,j is the susceptance of the internal electromotive force of generator j and the fault node i.

[0108] The risk of faults in the DC power transmission system leads to the breakdown of the external power balance and further causes power changes. At this time, the generator needs to spontaneously release kinetic energy to provide active support for the system. The specific support expression is as follows:

[0109]

[0110] In the formula: P h is the output power of active support; f(t) is the synchronous generator frequency.

[0111] If the frequency deviation value is larger, the support power output by the generator is higher. The relationship between the support power and the frequency change is as follows:

[0112]

[0113] In the formula: G av is the generator inertia response constant; f n is the rated frequency of the generator; f e is the extreme frequency point of the generator.

[0114]

[0115] In the formula: A i is the reduction factor in the matrix of node i; G min , G max are respectively the minimum and maximum values of the generator inertia constant.

[0116] The inertia parameter plays an important supporting role in the system operation. As a risk adjustment factor, it can significantly reduce the comprehensive outage probability of the system. This supporting ability is reflected through the negative correlation between inertia and system risk, thus enhancing the operation stability of the system.

[0117] p = χp u (9)

[0118] In the formula: p is the comprehensive outage model combined with the inertia constant; p u is the line outage probability model; χ is the reduction factor matrix, and its expression is:

[0119]

[0120] Research shows that there is a significant correlation between the node inertia and the system's anti-disturbance ability. Due to their inherent inertia characteristics, high-inertia nodes can maintain a relatively stable operating state under the action of disturbances. In contrast, due to their smaller inertia, low-inertia nodes are more likely to have their operating states significantly changed by external disturbances.

[0121] 2. System outage probability distribution entropy value index

[0122] The current research proposes to use the power flow entropy index to identify key lines. The load rate and its distribution are considered to characterize the importance of the line in the system. Due to the change of the power grid operation state, the load rate continues to change. If the load change rate w i ≤L ie / L imax When L , line i always operates in a normal state, but the power flow entropy index still changes with the load rate; ie / L imax <w i When <1, the grid operation condition will change. At this time, although the line flow will change accordingly, the positive correlation between its load rate and flow is different from the actual operation law of the line. Obviously, the above situation does not conform to the actual operation of the grid, so a probability distribution entropy index is proposed to characterize the relationship between line outage and grid operation status.

[0123] Let K 1 ,K 2 …K n is a set of constant sequences. When the target line fails, the outage probability of the i-th line is P i , for a given constant sequence K = {K 1 ,K 2 …K n}, Definition I k To satisfy P i ∈(K k ,K k+1 ) lines. Then the line outage probability in the system falls within the interval (K k ,K k+1 ) with a probability T k It can be expressed as:

[0124]

[0125] Assume that the outage probability distribution vector is T = [T 1 ,T 2 ,…T n-1 ]. If in the interval (K k ,K k+1 ) contains S k lines, and P i,k Indicates that when P i ∈(K k ,K k+1 ) is the outage probability of the ith line at the time, then the average outage probability of the line in this interval is W k It can be expressed as:

[0126]

[0127] Where: P i,k To satisfy Pi ∈(K k , K k+1 ), the outage probability of the i-th line

[0128] The entropy value G of the system outage probability distribution x is as follows:

[0129]

[0130] G x characterizes the degree of dispersion of the system outage probability distribution after the failure of the target line. When G x has a large value, it indicates that the system is more likely to trigger cascading failures under disturbances.

[0131] 3. Entropy index of system outage probability change

[0132] The importance of lines in the power grid is positively correlated with the intensity of the cascading effect caused by their failures. When the failure of a certain line only causes limited disturbances to other lines, its threat level to the overall operation stability of the system is relatively low. The existing power flow transfer entropy index has the limitation of not considering the overall load rate of the system. The identification of key lines in traditional models may face various situations that do not conform to actual operation. This patent establishes a new type of outage probability change entropy index based on the line outage probability model considering inertia.

[0133] Assume that after the failure of line i, the outage probability of other line k is denoted as By comparing the initial steady-state outage probability of the system with the new steady-state probability after the failure the outage probability when reaching the new steady state after obtaining the change in outage probability can be expressed as:

[0134]

[0135] Then the total change in the system outage probability is:

[0136]

[0137] In the formula: L represents the total number of lines in the power grid topology.

[0138] Define the proportion η k,i of the change in the outage probability of line k as:

[0139]

[0140] In the formula: ΔP t,i represents the total change in the system outage probability. Thus, the outage probability change vector η = [η 1,i , η 2,i , …, η L,i, and then define the entropy change of outage probability G c as follows:

[0141]

[0142] G c is used to quantify the discreteness of the system outage probability distribution after a fault occurs on the target line. When the value of G c is small, it indicates that the impact of the fault is concentrated on a few lines, and the system is more likely to experience cascading failures.

[0143] 4. Improved power grid topology entropy index

[0144] The identification of key lines in the power grid needs to comprehensively consider the topological structure characteristics. Based on the node importance evaluation method proposed by scholars, this patent takes into account the impact of line faults on the topological integrity of the power grid and combines the topological connection degree theory to construct an improved power grid structure entropy index.

[0145] Define the degree of node n as:

[0146]

[0147] where the node connection status quantity v a,b satisfies:

[0148]

[0149] In the formula: N is the total number of system nodes.

[0150] Given a constant sequence Q = {Q 1 , Q 2 , …, Q m}, define M k as the number of nodes whose degree R n ∈ (Q k , Q k+1 ), then the node degree distribution probability S k is:

[0151]

[0152] Thus, the node degree distribution vector S = [S 1 , S 2 , …, S m-1 can be obtained, and then define the improved entropy G f of line i as:

[0153]

[0154] where the topological connection degree H i is calculated as follows:

[0155]

[0156] Where: T k is the degree R n ∈(Q k , Q k+1 ) is the average node degree, and K i is the sum of the degrees of the two end nodes of line i. This index G f has the following characteristics:

[0157] 1) It reflects both the hub role of the line in power transmission and its impact on the topological integrity of the system after a fault.

[0158] 2) The larger the value, the higher the importance of the line in the power grid.

[0159] 3) It comprehensively considers the self-organized criticality and local connection ability of the power grid and can effectively evaluate the global and local influence of the line.

[0160] 5. Improved two-level grey critical line comprehensive evaluation method

[0161] The operational safety of the power grid is comprehensively affected by the overall operating state of the system, the operating state of local lines, and the stability of the topological structure. To comprehensively evaluate the criticality level of transmission lines in the power grid, a multi-dimensional comprehensive evaluation system needs to be established. This study proposes a comprehensive index model for critical lines that integrates inertia and multi-source information, specifically defined as representing the degree of balance of the line outage probability distribution, where G x is the calculated value of information entropy; reflecting the robust characteristics of the power grid topological structure, G f is the network structure entropy parameter; quantifying the discrete characteristics of the time-series change of the outage probability, G c is the entropy value of probability change.

[0162] For the multi-dimensional risk characteristics of the new power system, the structure of the improved two-level grey critical line comprehensive evaluation method is as Figure 2 shown.

[0163] The structure of the improved two-level grey critical line comprehensive evaluation method described above includes a criterion layer and an index layer. The three indexes of the system outage probability distribution entropy G x , the outage probability change entropy G c , and the improved power grid topology entropy G f are used to evaluate critical lines in two layers;

[0164] In the criterion layer, two core evaluation criteria are established to reflect the overall stability of the power grid during operation and the robustness of the structure. These two criteria are respectively:

[0165] 1) Operational stability C 1Layer: Evaluate the stability of the power grid under normal and disturbed conditions. This criterion focuses on the system's response and recovery capabilities in the face of emergencies such as faults and load changes, ensuring the continuous and safe operation of the power system;

[0166] 2) Structural robustness C 2 Layer: Evaluate whether the topological structure of the power grid can effectively resist potential faults and interferences; this criterion considers the rationality of the power grid configuration, the connections between nodes, and their support for system stability, avoiding widespread impacts caused by single-point failures.

[0167] The algorithm flow of the improved two-level grey key line comprehensive evaluation method is as follows:

[0168] Propose a sliding window normalization method to process time-varying data:

[0169]

[0170] In the formula: τ = 24h is the time window length, effectively capturing the intraday fluctuation characteristics.

[0171] Fuse the improved entropy weight method and dynamic AHP:

[0172]

[0173] In the formula: w j (t) represents the combined weight of the j-th index at time t; γ is the combined adjustment coefficient (0 ≤ γ ≤ 1) used to balance the ratio of the objective entropy weight and the subjective AHP weight.

[0174] Based on the objective weight calculated by the improved sliding entropy weight method, it reflects the information dispersion degree of the index data within the time window [t - τ, t]:

[0175]

[0176] In the formula: where E j (t) is the sliding information entropy, calculated by normalizing the data Calculate; p ij (t) represents the contribution degree of the i-th sample in the j-th index; t ∈ + Reflects the time characteristics of dynamic evaluation; Is the dynamic weight calculated by the entropy weight method; N is the total number of samples within the sliding time window. For every 10% increase in the new energy proportion, N decreases by 2. Using the optimal values of each index to form the reference sequence X 0 (t), calculate the correlation coefficient γ ij (t) and the correlation degree R i (t).

[0177] When using dynamic grey relational analysis, first construct a time-varying reference sequence:

[0178] X 0 G(t)=[minG x (t),minG c (t),maxG f (t)] (27)

[0179] Improved calculation of correlation coefficient:

[0180]

[0181] Among them, the discrimination coefficient ρ(t) is adaptively adjusted:

[0182]

[0183] Establish a two-layer correlation degree synthesis model:

[0184]

[0185] In the formula: Δ ij G(t) is the absolute difference between the i-th object and the reference sequence in the j-th index; Δ min All Δ ij The minimum value in G(t); Δ max All Δ ij The maximum value in G(t); Var(X(t)) is the coefficient of variation (standard deviation / mean) of the index matrix, w c G(t) is the weight of the criterion layer (c = 1 for operation stability, c = 2 for structural robustness); J c The index set corresponding to the criterion layer c; where ρ = 0.5 is the discrimination coefficient, and R i (t) The larger the value, the higher the criticality of the line. The algorithm implementation process is as Figure 3 shown.

[0186] This model innovatively integrates four dimensions: the low-inertia characteristics of high-proportion new energy nodes, static probability distribution, dynamic evolution characteristics, and network structure characteristics. Through the entropy value theory, the parameters with different dimensions are normalized, overcoming the limitations of traditional single-index evaluation. Its overall identification process is as Figure 4 shown.

[0187] The specific steps for identifying the critical line are as follows:

[0188] 1) Establish an initial model of the DC external power transmission system, extract the initial topological structure of the power grid from the database, including nodes (substations, generators) and lines (transmission lines), and calculate the initial outage probability distribution vector T according to formula (11). This vector reflects the fault risk of each line in the initial state.

[0189] 2) Restore the lines that were disconnected due to faults or maintenance before to obtain the complete grid operation status. Determine the next target line to be identified, de-energize this line for analysis, while keeping the other parts of the grid operating normally.

[0190] 3) Based on step 2, calculate the power flow and node degree distribution of each line after de-energization (the degree refers to the number of lines connected to a certain node and the importance of the node). Using these data, calculate the line outage probability distribution vector T, the line outage probability change vector η, and the node degree distribution vector S after de-energization.

[0191] 4) Calculate the indexes G x , G c and G f respectively according to formulas (13), (17), and (21). Check whether all valid lines have been traversed. If all lines have been processed, go to step 5; if there are still untraversed lines, return to step 2 and continue to analyze the next line.

[0192] 5) Conduct an improved two-level grey critical line evaluation for each index G x , G c and G f to obtain the comprehensive index of the critical line.

[0193] 6) Arrange the comprehensive indexes of the critical lines in descending order according to their values to identify the critical lines in the grid.

[0194] Experiment of the present invention:

[0195] Build a test scenario with a 30% wind power penetration rate for the new energy UHV DC external power transmission system based on PSCAD, and generate 2000 groups of operation samples. The calculation results of the index weights show that w x = 0.318, w f = 0.325, w c = 0.357, indicating that the entropy of the system outage probability change has the most significant impact on the criticality assessment. It can be seen from Table 1 that the coincidence degree between the top 5 lines in the grey correlation degree ranking and the traditional N-1 check results reaches 89.3%, and the hidden critical lines caused by insufficient inertia support (such as Line8-9, Line 16-17) are accurately identified. It can be seen from Table 2 that compared with the single-index evaluation, the identification results of the present invention are more accurate and have better robustness and stability.

[0196] Table 1 Comparison of critical line identification results

[0197]

[0198] Table 2 Comparison of the performance of identification methods

[0199]

Claims

1. A method for identifying key lines of a large-scale new energy DC transmission system considering inertia, characterized in that: This identification method uses the supporting relationship between node inertia and system stability to introduce the inertia risk adjustment factor to establish a system outage probability model considering the spatiotemporal distribution of node inertia; secondly, on the basis of this model, the uniformity and degree of the system outage probability distribution after the target line is shut down are analyzed, and combined with the global topological structure of the system, a three-dimensional identification index system is constructed, including the system outage probability distribution entropy index, the system outage probability change entropy index and the improved power grid topology entropy index; finally, the three-dimensional identification index results are generated by reading the initial topological structure and operating parameters of the power grid, the entropy weight method is used to determine the weight of each index, and the improved two-level grey key line comprehensive evaluation method is used to synthesize the three-dimensional identification index, so as to identify the key lines and sort the risks.

2. According to the method for identifying key lines of a large-scale new energy DC transmission system considering inertia according to claim 1, it is characterized in that: The establishment of the system outage probability model considering the spatiotemporal distribution of node inertia is as follows: first, the line outage probability model is established by considering the dynamic operation characteristics of the system: including the influence of line flow, bus voltage and system frequency and potential hidden faults, and its expression is as follows: Where: p u is the line outage probability; P L Reflects the probability of hidden faults in the relay protection system; P H Describes the risk of line overload; in the model parameters, L i represents the real-time power flow of line i; L ie is the normal operating limit of the line; L imax Represents the maximum transmission power of the line; The concept of system inertia is introduced into the above outage probability model, and the expression of the total inertia constant of the system is: Where: G S is the total inertia constant of the system; m is the number of generators in the DC transmission system; G j is the inertia constant of generator j; There is a specific correlation between the inertia constant of each node and the disturbance amplitude: Where: G i is the inertia constant corresponding to the system node i during disturbance; f z is the system node frequency; Y i,j is the voltage correlation matrix between node i and generator j; B i,j is the potential inside generator j and the susceptance of fault node i; The risk of failure in the DC transmission system causes the external power balance to break and then causes power changes. At this time, the generator needs to spontaneously release kinetic energy to provide active support for the system. The specific support expression is: Where: P h is the active support output power; f(t) is the synchronous generator frequency; If the frequency deviation value is larger, the support power output by the generator will be higher. The relationship between the support power and the frequency change is: Where: G av is the generator inertia response constant; f n is the rated frequency of the generator; f e is the extreme point of generator frequency; Where: A i is the reduction factor in the matrix of node i; G min , G max are the minimum and maximum values ​​of the generator inertia constant respectively; Inertia parameters play an important supporting role in system operation. As a risk adjustment factor, they can significantly reduce the overall probability of system outage. This support capacity is reflected by the negative correlation between inertia and system risk: p = χp u (9) Where: p is the comprehensive outage model combined with the inertia constant; p u is the line outage probability model; χ is the reduction factor matrix, and its expression is:

3. According to the method for identifying key lines of a large-scale new energy DC transmission system considering inertia in claim 1, it is characterized in that: The system outage probability distribution entropy index is used to characterize the relationship between line outage and power grid operation status; it is set as follows: Let K1, K2…K n is a set of constant sequences; when the target line fails, the outage probability of the i-th line is P i , for a given constant sequence K = {K1, K2…K n }, Definition I k To satisfy P i ∈(K k ,K k+1 ) of the lines; then the probability of line outage in the system falls within the interval (K k ,K k+1 ) with a probability T k It can be expressed as: Assume that the outage probability distribution vector is T = [T1, T2, ... T n-1 ]; if in the interval (K k ,K k+1 ) contains S k lines, and P i,k Indicates that when P i ∈(K k ,K k+1 ) is the outage probability of the ith line at the time, then the average outage probability of the line in this interval is W k It can be expressed as: Where: P i,k To satisfy P i ∈(K k ,K k+1 )’s outage probability of the ith line; System outage probability distribution entropy value G x for: G x It represents the discrete degree of the probability distribution of system outage after the target line fails. x When the value is large, it indicates that the system is more likely to cause cascading failures under disturbance.

4. According to the method for identifying key lines of a large-scale new energy DC transmission system considering inertia according to claim 1, it is characterized in that: The system outage probability change entropy index is used to quantify the discrete degree of the system outage probability distribution after the target line fails, and is set as follows: Assuming that line i fails, the outage probability of other lines k is recorded as By comparing the system steady-state initial outage probability and the new steady-state probability of failure The change in outage probability can be obtained The expression is: The total change of system outage probability is: Where: L represents the total number of lines in the power grid topology; Define the proportion of the change in the outage probability of line k as η k,i for: Where: ΔP t,i represents the total system outage probability change; thus, the outage probability change vector η=[η 1,i ,η 2,i ,…,η L,i ], and then define the outage probability change entropy G c for: When G c When the value is small, it indicates that the fault impact is concentrated on a few lines and the system is more likely to have cascading failures.

5. According to the method for identifying key lines of a large-scale new energy DC transmission system considering inertia in claim 1, it is characterized in that: The improved power grid topology entropy index is set as follows: Define the degree of node n as: Among them, the node connection state v a,b satisfy: Where: N is the total number of system nodes; Given a constant sequence Q = {Q1, Q2, ..., Q m }, define M k R n ∈(Q k ,Q k+1 ) of the number of nodes, then the node degree distribution probability S k for: Thus we can get the node degree distribution vector S = [S1, S2, ..., S m-1 ], and then define the improved entropy G of line i f for: Among them, the topological connection degree H i The calculation is as follows: Where: T k R n ∈(Q k ,Q k+1 )’s average node degree, K i is the sum of the degrees of the nodes at both ends of line i; G f It reflects the robustness of the power grid topology. The larger the value, the more important the line is in the power grid.

6. The key line identification method of a large-scale new energy DC transmission system considering inertia according to claim 1 is characterized in that: The structure of the improved two-level grey key line comprehensive evaluation method includes a criterion layer and an indicator layer. x , Outage probability change entropy G c , Improved grid topology entropy G f The three indicators are divided into two layers to evaluate key routes; In the criterion layer, two core evaluation criteria are established to reflect the overall stability of the power grid and the robustness of the structure during operation. These two criteria are: 1) Operational stability C1 layer: evaluates the stability of the power grid under normal and disturbance conditions; This criterion focuses on the system's ability to respond and recover in the face of emergencies, such as failures and load changes, to ensure that the power system can maintain continuous and safe operation; 2) Structural robustness C2 layer: evaluates whether the topology of the power grid can effectively resist potential faults and interference; This criterion considers the rationality of the grid configuration, the connections between nodes and their support for system stability to avoid widespread impacts due to single point failures.

7. According to claim 6, a double-layer evaluation criteria method for key lines of a large-scale new energy DC transmission system considering inertia is characterized in that: Using the double-level grey key line comprehensive evaluation method, the algorithm first proposes a sliding window standardization method to process time-varying data: Where: τ = 24h is the time window length, which effectively captures the intraday volatility characteristics; Fusion of improved entropy weight method and dynamic AHP: Where: w j (t) represents the combined weight of the jth indicator at time t; γ is the combined adjustment coefficient (0≤γ≤1) used to balance the ratio of objective entropy weight to subjective AHP weight; The objective weight calculated based on the improved sliding entropy weight method reflects the degree of information dispersion of the indicator data in the time window [t-τ, t]: In the formula: E j (t) is the sliding information entropy, through the standardized data Calculated; p ij (t) represents the contribution of the i-th sample to the j-th indicator; t∈ + Reflect the temporal characteristics of dynamic evaluation; is the dynamic weight calculated by the entropy weight method; N is the total number of samples in the sliding time window. When the proportion of new energy increases by 10%, N decreases by 2. The reference sequence X0(t) is constructed with the optimal values ​​of each indicator, and the correlation coefficient γ is calculated. ij (t) and the correlation R i (t); When using dynamic grey relational analysis, first construct a time-varying reference sequence: X0(t)=[minG x (t),minG c (t),maxG f (t)] (27) Improved correlation coefficient calculation: The resolution coefficient ρ(t) is adaptively adjusted: Establish a two-layer correlation synthesis model: Where: Δ ij (t) is the absolute difference between the i-th object and the reference sequence in the j-th index; Δ min All Δ ij (t) minimum value; Δ max All Δ ij The maximum value in (t); the coefficient of variation (standard deviation / mean) of the indicator matrix Var(X(t)) w c (t) Criterion layer weight (c = 1 for operational stability, c = 2 for structural robustness); J c The set of indicators corresponding to the criterion layer c; where ρ = 0.5 is the resolution coefficient, R i The larger the (t) value, the higher the line criticality.

Images