A Transmission Network Cascading Fault Risk Assessment Device and Method Based on an Optimized Domain

Abstract

The present invention discloses a transmission network cascading failure risk assessment device and method based on an optimization domain, belonging to the technical field of transmission network fault assessment, including: a cascading failure optimization domain generation module for constructing a cascading failure risk assessment model based on historical data and generating an optimization domain; a network disconnection determination module for judging the on-network status of all generators in the transmission network based on the current operating state; a cascading failure deduction module for determining cascading line failures caused by a single fault and updating the optimization domain based on the current topological state; a cascading failure assessment module for constructing a mapping from the operating space to the optimization domain based on all potential renewable energy scenarios to achieve rapid cascading failure assessment. By adopting the above-mentioned transmission network cascading failure risk assessment device and method based on an optimization domain, the present invention can effectively evaluate the impact of large-scale scenarios on cascading failures and can significantly improve the speed and efficiency of cascading failure assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of power transmission network fault assessment, and in particular to a device and method for assessing the risk of cascading faults in a power transmission network based on an optimization domain. Background Technique

[0002] In recent years, the frequent occurrence of large-scale power outages has posed a major threat to the safe and stable operation of power systems. For example, the power outage event of the Western Interconnection Grid in July 1996, the major power outage in Argentina in 2019, and the major power outage in Mexico in 2020 all indicate that the failures of multiple components within the system may trigger cascading faults, thereby resulting in large-scale power outage losses. Therefore, it is very necessary to consider the risk of potential cascading faults during the operation stage.

[0003] Sequential Monte Carlo simulation and the finite-state continuous-time Markov chain method are widely used to assess cascading faults. However, considering a large number of potential operating scenarios of renewable energy in these assessments will lead to the problem of combinatorial explosion. Therefore, these methods are difficult to apply in real-time operations. To solve this problem, a data mining-based model has been proposed to identify the fault propagation patterns in power systems.

[0004] However, these methods require a large amount of actual data for fitting and mining, and the risk assessment and fault evolution models have a great dependence on the data quality. In addition, artificial intelligence models have also been widely applied to the assessment of cascading faults in power systems. In the prior art, a cascading fault assessment method based on graph convolutional networks has been proposed, which uses graph convolutional neural networks and long short-term memory networks to extract the electrical characteristics of the system from the spatial and temporal dimensions respectively. However, artificial intelligence models are usually black-box models and are difficult to apply in real-time operations. To sum up, assessing the risk of cascading faults usually requires evaluating a large number of fault propagation paths and a large number of renewable energy scenarios. Summary of the Invention

[0005] The purpose of the present invention is to provide a device and method for assessing the risk of cascading faults in a power transmission network based on an optimization domain, which can avoid the assessment errors caused by the pre-reduction of scenarios in traditional methods, thereby being able to effectively evaluate the impact of large-scale scenarios on cascading faults, and can dynamically identify the optimal scenario risk assessment only by simply determining the position of the optimization domain where the current operating state is located, which can significantly improve the speed and efficiency of cascading fault assessment.

[0006] To achieve the above object, the present invention provides a device for assessing the risk of cascading faults in a power transmission network based on an optimization domain, including:

[0007] A cascading fault optimization domain generation module, configured to construct a cascading fault risk assessment model based on historical data and generate an optimization domain;

[0008] The off-grid determination module is used to judge the on-grid status of all generators in the transmission network based on the current operating state;

[0009] The cascading fault deduction module is used to determine the cascading line faults caused by a single fault and update and optimize the domain based on the current topological state;

[0010] The cascading fault assessment module is used to construct the mapping from the operating space to the optimization domain based on all potential renewable energy scenarios to achieve rapid cascading fault assessment.

[0011] The present invention also provides a method for assessing the risk of cascading faults in a transmission network based on an optimization domain, including the following steps:

[0012] S1. By obtaining the operating state of the power transmission system, a corresponding cascading fault analysis model is constructed;

[0013] S2. An optimization domain is established in advance R A , and all possible operating scenarios are mapped to points in the corresponding optimization domain through a linear space;

[0014] S3. A generator off-grid model based on probability is established to sense the operating state of the generator and dynamically determine whether the generator set should be off-grid;

[0015] S4. Sense the power grid topology;

[0016] S5. The cascading fault analysis model is compactly formulated by matrix representation to obtain a cascading fault assessment model;

[0017] S6. The cascading fault assessment model in S5 is further corrected based on S3 and S4;

[0018] S7. Perform optimization domain division and operating state matching;

[0019] S8. Obtain the optimal solution through the corresponding basis items.

[0020] Preferably, in S1, the cascading fault analysis model is specifically:

[0021] ;

[0022] ;

[0023] ;

[0024] ;

[0025] ;

[0026] Where: is the optimal solution of the objective function; is the cost vector of load loss; is the load loss value; represents the nodal admittance matrix; represents the phase angle; and respectively represent the generator output and the load demand; represents the power flow of branch i and j; and are the upper and lower limits of branch power respectively; is the upper limit of the generator output.

[0027] Preferably, in S2, the optimization domain has the expression:

[0028] ;

[0029] Where: are several regions that make up the optimization domain, E is the identity matrix; is the cost vector of load loss corresponding non-basic variables; is the cost vector of load loss corresponding basic variables; B , N are the basic corresponding terms and the non-basic corresponding terms respectively, diag( D , E ) represents a diagonal matrix with D , E as the diagonal elements.

[0030] Preferably, in S3, the expression of the established generator tripping model is:

[0031] ;

[0032] Where: is the type of renewable energy, including photovoltaic and wind turbines; is the device on the grid probability; represents the voltage magnitude of device k; and represent the low voltage disconnection threshold voltage of device k; and represent the high voltage disconnection threshold voltage of device .

[0033] Preferably, in S4, a current threshold is preset. When the branch power flow exceeds the threshold, the protection device will trip and disconnect the line, and the cascading fault deduction model is determined:

[0034] ;

[0035] Wherein: represents the fault probability of the branch , represents the threshold value at which the protection device on the branch is activated;

[0036] During the cascading fault process, only one branch trip is allowed each time. When at least two branches are overloaded simultaneously, the branch with the most severe overload is selected in sequence for trip processing;

[0037] When the fault causes cascading trips in the line, it will change the topology of the power transmission network, and thus change the nodal admittance matrix , therefore, in the cascading fault analysis model in S1 is modified to: , wherein: is the topology change matrix; the corresponding optimization domain is updated to:

[0038] .

[0039] Preferably, the expression of the cascading fault assessment model in S5 is:

[0040] .

[0041] Preferably, in S6, if it is determined based on S3 and S4 that the injection power and network topology have changed, the cascading fault assessment model is modified to:

[0042] ;

[0043] Wherein: and are the generator output fluctuation vector and the load fluctuation vector respectively;

[0044] According to the linear programming theory, the cascading fault assessment model is expressed in the following form of basis B and non-basis N:

[0045] .

[0046] Preferably, in S7, any and fluctuations corresponding operating states correspond to an optimization point in the optimization domain RA , and the optimization points with the same operating characteristics together form a subset in the optimization domain, determined by the basis parameters B and non-basis parameters N;

[0047] For any operating state and , the matching relationship between the optimization domain and the operating state is determined by the following formula:

[0048] ;

[0049] .

[0050] Preferably, in S8, the optimal solution is calculated by the formula:

[0051] .

[0052] Therefore, the beneficial effects of the present invention adopting the above-mentioned power transmission network cascading fault risk assessment device and method based on the optimization domain are as follows:

[0053] (1) Optimize the computing and processing capabilities in complex scenarios: The proposed cascading fault analysis model based on the optimization domain transforms complex optimization problems into linear equation solving problems, and significantly improves the computing efficiency by constructing the operating relationships between various operating states and the optimization domain.

[0054] (2) Improve the efficiency of cascading fault assessment: The present invention can effectively analyze voltage limit and branch power flow overload problems, especially in the case where renewable energy generators are disconnected due to voltage problems, and quickly evaluate the system. Compared with the prior art, in the same scenario, the evaluation time of the present invention is reduced to 12.28% of the prior art, effectively reducing the evaluation time of the subsequent impact of faults.

[0055] (3) Strong adaptability and acceleration effect: Since the cascading fault deduction module of the present invention can dynamically update the optimization domain corresponding to the topological state, therefore, the computing efficiency of the present invention is independent of the topological type, load type, and cascading fault propagation path, with strong adaptability, ensuring the universality of the acceleration efficiency.

[0056] Next, through the drawings and embodiments, the technical solutions of the present invention will be further described in detail. Brief Description of the Drawings

[0057] Figure 1 is a schematic diagram of the composition of a power transmission network cascading fault risk assessment device based on the optimization domain of the present invention;

[0058] Figure 2 is a generator tripping model diagram of an embodiment of a power transmission network cascading fault risk assessment method based on the optimization domain of the present invention;

[0059] Figure 3 is a comparison diagram of the calculation time of a power transmission network cascading fault risk assessment method based on the optimization domain of the present invention and the prior art based on the optimization domain. Detailed Embodiments

[0060] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0061] Embodiment 1

[0062] As Figure 1 shown, the present invention provides a transmission grid cascading failure risk assessment device based on an optimization domain, including:

[0063] A cascading failure optimization domain generation module for constructing a cascading failure risk assessment model based on historical data and generating an optimization domain.

[0064] A network disconnection determination module for judging the on-network status of all generators in the transmission grid based on the current operating state.

[0065] A cascading failure deduction module for determining cascading line failures caused by a single failure and updating the optimization domain based on the current topological state.

[0066] A cascading failure assessment module for constructing a mapping from the operating space to the optimization domain based on all potential renewable energy scenarios to achieve rapid cascading failure assessment.

[0067] The present invention also provides a transmission grid cascading failure risk assessment method based on an optimization domain, including the following steps:

[0068] S1. By obtaining the operating state of the power transmission system, a corresponding cascading failure analysis model is constructed.

[0069] The cascading failure optimization domain generation module is based on the operating state variables of the historical power transmission network, which include: information such as the voltage, active load, and reactive load of nodes; the active output and reactive output of generators and renewable energy units; and the active power flow and reactive power flow of lines.

[0070] By obtaining the above information, a corresponding cascading failure analysis model is further constructed, specifically:

[0071]

[0072]

[0073]

[0074]

[0075]

[0076] Wherein: is the optimal solution of the objective function; is the cost vector of load loss; is the load loss value; represents the nodal admittance matrix; represents the phase angle; and represent the generator output and load demand respectively; represents the power flow of branch i and j; and are the upper and lower limits of branch power respectively; is the upper limit of the generator output.

[0077] S2. Pre - establish the optimization domain , and map all possible operating scenarios to points in the corresponding optimization domain through a linear space.

[0078] The optimization domain is defined as the optimization variable space generated by the parameter space of the cascading failure analysis model. Specifically, since the operating states such as the output of renewable energy units, load demand, and generator output change continuously over time, it is necessary to frequently solve the cascading failure analysis model, which will consume a large amount of time in actual cascading failure assessment. Considering that cascading failures usually occur on an extremely short - time scale, traditional solution methods cannot meet the requirements of rapid response. Therefore, in this embodiment, by pre - defining the optimization domain, all possible operating scenarios are mapped to points in the corresponding optimization domain through a linear space. For any cascading failure state, the system only needs to determine the optimization domain where the state point is located, and then can quickly complete the analysis and assessment of cascading failures, greatly improving the fault response speed and assessment efficiency, and ensuring that the power system can take countermeasures in time when cascading failures occur to avoid further expansion of the failures.

[0079] Optimization domain R A The expression of is:

[0080] ;

[0081] Where: are several regions that make up the optimization domain, E is the identity matrix; is the cost vector of load loss corresponding non - basic variables; is the cost vector of load loss corresponding basic variables; B , N are the basic - corresponding term and non - basic - corresponding term respectively, diag( D , E ) represents a diagonal matrix with D , E as diagonal elements.

[0082] S3. Establish a probability-based generator disconnection model to sense the operating state of the generator and dynamically determine whether the generator set should be disconnected from the grid.

[0083] The disconnection determination module works based on the sensing results of the cascading failure optimization domain generation module. When the voltage at the node where the generator set or renewable energy generator set is connected exceeds the preset safety threshold, this module triggers the protection mechanism of the generator set, thereby realizing the disconnection operation of the generator set. The disconnection determination module dynamically determines whether the generator set should be disconnected from the grid by establishing a probability-based generator disconnection model, as Figure 2 shown, comprehensively considering multi-dimensional factors such as the frequency, amplitude, and duration of voltage fluctuations. Thus, it ensures the safe and stable operation of the power system and avoids the further expansion of voltage abnormalities affecting the overall stability of the system.

[0084] The expression of the established generator disconnection model is:

[0085] ;

[0086] Where: is the type of renewable energy, including photovoltaic and wind turbines; is the probability of the device on the grid; represents the voltage amplitude of device k; and represent the low voltage disconnection threshold voltage of device k; and represent the high voltage disconnection threshold voltage of device .

[0087] S4. Sense the grid topology.

[0088] In the cascading failure deduction module, a current threshold is preset in advance. When the branch power flow exceeds the threshold, the protection device will trip and disconnect the line, and the cascading failure deduction model is determined:

[0089] ;

[0090] Where: represents the failure probability of branch , represents the threshold at which the protection device on branch is activated.

[0091] In this embodiment, it is set that during the cascading fault process, only one branch is allowed to trip each time. When overload occurs in at least two branches simultaneously, the branch with the most severe overload is selected in sequence for tripping. In addition, the time interval between different tripping lines is considered to be extremely short. However, considering that the cascading fault assessment module of this embodiment has an efficient state analysis ability and can update and analyze the system state in real time within an extremely short time, it is reasonable and feasible to evaluate scenarios for multiple branches within an extremely short time interval.

[0092] When a fault causes cascading tripping of lines, it will change the topology of the power transmission network, and thus change the nodal admittance matrix , therefore, in the cascading fault analysis model is modified to: , where: is the topology change matrix.

[0093] The corresponding optimization domain is updated to:

[0094] .

[0095] S5. Compact the formulation of the cascading fault analysis model by matrix representation method to obtain the cascading fault assessment model.

[0096] The expression of the cascading fault assessment model is:

[0097] .

[0098] S6. Further correct the cascading fault assessment model in S5 based on S3 and S4.

[0099] At each stage of the cascading fault evolution, the network topology changes due to line overload, and the changes in power injection caused by device disconnection are respectively reflected as changes in the parameters , and . Therefore, determine whether the injection power and topology change through the generator tripping determination model in S3 and the cascading fault deduction model in S4. If they change, the cascading fault assessment model obtained in S5 is corrected in the following general form:

[0100] ;

[0101] where: and are the generator output fluctuation vector and the load fluctuation vector respectively. According to the linear programming theory, the cascading fault assessment model is expressed in the following form of basis (B) and non - basis (N):

[0102]

[0103] S7. Optimize domain division and operation status matching.

[0104] any and The operating state corresponding to the fluctuation corresponds to the optimization domain RA An optimization point in the optimization domain, and the optimization points with the same operating characteristics together constitute a subset in the optimization domain, which is determined by the basic parameter B and the non-basic parameter N.

[0105] For any operating state and , the matching relationship between the optimization domain and the operating state is determined by the following formula:

[0106] ;

[0107] .

[0108] S8. Obtain the optimal solution through basis correspondences.

[0109] Points in the same optimization domain RA have the same basis correspondences (B, N), and the optimal solution can be obtained directly through the basis correspondences without performing optimization calculations.

[0110] Optimal solution The calculation formula is:

[0111] .

[0112] In this embodiment, the test system adopts the IEEE 39 bus system, including 34 transmission lines, 12 transformers and 10 generators. In addition, wind turbines and photovoltaic power stations are added, and the specific capacities are shown in Table 1.

[0113] Two schemes are set up for comparison: (1) risk assessment considering random failures; (2) comprehensive risk assessment considering cascading failures.

[0114] Table 1 Parameters of wind turbines and photovoltaic power stations

[0115]

[0116] Table 2 shows the safety analysis results considering random faults and cascading faults. It can be seen that scenario 1 only considers the impact of the main fault branch on the system, and does not consider the potential multi-stage fault cascade process that may be triggered by the main fault. Therefore, compared with scenario 2, the safety analysis results of scenario 1 are more conservative. For example, in the case of a major fault in three devices, the load shedding in scenario 1 is only 48.94% of that in scenario 2.

[0117] Comparison Results of IEEE 39-Bus System

[0118]

[0119] In addition, through Figure 3 It can be seen that the calculation time of the cascading failure assessment model based on the optimization domain is only 12.28% of that of the traditional method. This is because the optimization domain method converts most of the optimal power flow process into the generation and matching process of the optimization domain. However, since the operating states need to be classified in the optimization domain method, an additional computational burden is introduced, accounting for about 2.9%.

[0120] Therefore, the present invention adopts the above-mentioned device and method for risk assessment of cascading failures in a transmission network based on the optimization domain, which can avoid the evaluation error caused by the pre-reduction of scenarios in the traditional method, and thus can effectively evaluate the impact of large-scale scenarios on cascading failures. Only by simply determining the position of the optimization domain where the current operating state is located can the optimal scenario risk assessment be dynamically identified, which can significantly improve the speed and efficiency of cascading failure assessment.

[0121] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A transmission network cascading failure risk assessment method based on optimization domain, characterized in that: The following steps are involved: S1. By acquiring the operating state variables of the historical power transmission network, a corresponding cascading failure analysis model is constructed; In S1, the cascading failure analysis model is specifically: ； ； ； ； ； in: is the optimal solution of the objective function; is the cost vector of load loss; is the load loss value; represents the node admittance matrix; represents the phase angle; and Represent generator output and load demand respectively; represents the power flow of branches i and j; and They are the upper and lower limits of branch power respectively; is the upper limit of the generator’s output; S2. Pre-establish optimization domain , map all possible operation scenarios to points in the corresponding optimization domain through linear space; In S2, the optimization domain The expression is: ； in: are several regions that make up the optimization domain. E is the identity matrix; is the cost vector of load loss The corresponding non-basic variables; is the cost vector of load loss The corresponding base variable; B , N are the basis correspondence and non-basis correspondence respectively, diag( D , E ) indicates that D , E is a diagonal matrix with diagonal elements; S3. Establish a probability-based generator off-grid model to sense the operating status of the generator and dynamically determine whether the generator set should be off-grid; S4, sensing the topology of the power grid; In S4, the current threshold is preset. When the branch power flow exceeds the threshold, the protection device will trip and disconnect the line, and the cascading fault deduction model is determined: ； in: Indicates branch The probability of failure, Indicates branch The threshold at which the protective device is activated; During a cascading fault, only one branch is allowed to trip at a time. When at least two branches are overloaded at the same time, the branch with the most serious overload will be selected in turn for tripping. When a fault causes a chain tripping of the line, the topology of the transmission network will change, thereby changing the node admittance matrix. Therefore, the cascading failure analysis model in S1 was modified to: ,in: is the topology change matrix; the corresponding optimization domain Updated to: ； S5, compactly formulating the cascading failure analysis model through matrix representation to obtain a cascading failure assessment model; The expression of the cascading failure assessment model in S5 is: ； S6, further modifying the cascading failure assessment model in S5 based on S3 and S4; In S6, if the injected power and network topology are changed based on S3 and S4, the cascading failure assessment model is modified as follows: ； in: and are the generator output fluctuation vector and the load fluctuation vector respectively; According to linear programming theory, the cascading failure assessment model is expressed in the following form of base B and non-base N: ； S7, performing optimization domain division and operation status matching; In S7, any and The operating state corresponding to the fluctuation corresponds to the optimization domain RA An optimization point in , the optimization points with the same operating characteristics together constitute a subset in the optimization domain, which is determined by the basis corresponding item B and the non-basis corresponding item N; For any operating state and , the matching relationship between the optimization domain and the operating state is determined by the following formula: ； ； S8. Obtain the optimal solution of the cascading failure assessment model through basis correspondence items.

2. The method for assessing the risk of cascading failures in a power transmission network based on an optimization domain according to claim 1, characterized in that: In S3, the expression of the generator off-grid model is established as follows: ； in: types of renewable energy, including photovoltaics and wind turbines; Yes Equipment Probability on a grid; Indicates the device The voltage amplitude; and Indicates the device The low voltage disconnect threshold voltage; and Indicates the device The high voltage disconnect threshold voltage.

3. The method for assessing the risk of cascading failures in a power transmission network based on an optimization domain according to claim 1, characterized in that: In S8, the optimal solution The calculation formula is: 。 4. A device for a transmission network cascading failure risk assessment method based on an optimization domain as claimed in any one of claims 1 to 3, characterized in that: include: A cascading failure optimization domain generation module is used to build a cascading failure analysis model based on historical data and generate an optimization domain; An off-grid determination module is used to determine the on-grid status of all generators in the transmission network based on the current operating status; The cascading failure deduction module is used to determine the cascading line failures caused by a single failure and update the optimization domain based on the current topology state; The cascading failure assessment module is used to construct a mapping from the operating space to the optimization domain based on all potential renewable energy scenarios to achieve cascading failure assessment.

Images