Equipment cascading failure risk assessment method and system based on distribution network twinborn model

Abstract

The invention discloses an equipment cascading failure risk assessment method and system based on a distribution network twinborn model. The method comprises the following steps of: respectively generating an N-1-k anticipated accident set and an N-1-1 anticipated accident set when different initial anticipated shutdown elements fail according to a failure state of each equipment element and a topological relation among the equipment elements in a distribution network, wherein the N-1-k anticipated accident set and the N-1-I-1 anticipated accident set are generated when the initial anticipated shutdown elements fail; the method comprises the following steps: extracting a topology connection mode of a distribution network topology judgment initial anticipated shutdown element in a distribution network, and generating an N-1-k anticipated accident set or an N-1-1 anticipated accident set based on the topology connection mode; calculating the failure probability of each equipment element in the distribution network, and calculating the low-voltage risk of all nodes and the overload risk of all lines in the anticipated accident set for each anticipated accident set, so as to generate the distribution network cascading failure risk for each anticipated accident set; and sorting the importance of different anticipated accident sets based on the distribution network cascading failure risk of each anticipated accident set.

Description

Technical Field

[0001] The present invention relates to a field, and more specifically, to a method and system for assessing the risk of equipment cascading failures based on a distribution network twin model. Background Art

[0002] As the demand for electricity in distribution networks continues to rise year by year, electricity usage scenarios have changed from the previous single power supply mode to a diversified service mode. Distributed power sources, smart charging piles and other scenarios have put forward higher requirements for the digital transformation of distribution network services. As the "last mile" of serving customers, the distribution network needs to build a "strong distribution network and digital twin power grid" through digitalization, networking and intelligence to achieve full perception of the distribution network's "equipment status, operating environment, operation risks, and user electricity consumption".

[0003] The construction of information technology for power grid equipment management has gone through the SG186 project and the "SG-ERP" project stages. It has built and applied information systems such as PMS2.0, power supply service command, and intelligent operation and inspection control. It has actively explored and practiced new technologies such as the Internet of Things and artificial intelligence, and fully supported the work in equipment intrinsic safety, business lean management, and power supply service guarantee. At present, the application of information technology for power grid equipment management has insufficient capabilities in data models, system architecture, business integration, safety control, quality improvement and efficiency increase, high-quality services, and asset operation management. Building a new generation of equipment asset lean management system (PMS3.0), creating a digital architecture for a modern equipment management system, focusing on key areas, connecting key links, and solving outstanding problems have become the key points of the digital transformation of power grid equipment management.

[0004] The digital management of power grid equipment must focus on ensuring lean management and efficient operation of equipment assets, promote the sharing of data resources in all elements, all businesses, and the entire chain of asset management, realize the four digitalizations of "equipment, operation, management, and collaboration", enhance eight core business capabilities, and comprehensively improve the safety, quality, efficiency, and benefits of the company's equipment asset management.

[0005] Among them, equipment digitization promotes the deep integration of equipment with advanced sensing, information communication, automatic control and other technologies, establishes holographic digital portraits, supports the intelligence of main network equipment and the transparency of distribution network equipment, and improves the state perception ability and interconnection and interaction ability. Operation digitization integrates multi-dimensional perspectives such as on-site operations and team management, strengthens the construction of applications such as on-site analysis of equipment status, operation progress quality control, and digital engineering handover, realizes online, mobile, and transparent production operations, and improves on-site operation capabilities and safety control capabilities. Management digitization promotes data sharing, integrates equipment physical and value data, implements equipment risk prevention, health trend analysis, and asset performance evaluation, and improves risk prevention and control capabilities and precise investment capabilities. Collaborative digitalization integrates the digital multi-dimensional management system of various departments, establishes digital linkage applications with multiple entities such as users, suppliers, and service providers, builds an internal and external equipment asset management ecosystem, promotes the coordinated development of upstream and downstream industries, and improves management collaboration capabilities and value creation capabilities.

[0006] In the process of research and application of key technologies for digital twins of power grid equipment, the power grid resource business middle platform is the main data source of data twins and the interactive carrier for realizing twin services. The construction of the digital twin base of power grid equipment based on the power grid resource business middle platform can better achieve synchronization and consistency between virtual and physical data of power grid equipment based on the existing power grid data accumulation. "Digital twin" microservice applications that meet the technical architecture specifications of the power grid resource business middle platform can better integrate various applications with the new generation of equipment asset lean management system (PMS3.0).

[0007] However, there are still some problems in the application of distribution network digital twins. For example, the device perception capability is insufficient. At present, the online monitoring methods of some key equipment are not rich, and the distribution network perception capability is lacking. The supporting means for the real-time and data quality of IoT perception are insufficient. It is necessary to further enhance the device perception capability and improve the ability to aggregate data in real time and respond quickly to events at different levels, dimensions and regions. The integration degree of artificial intelligence algorithms is not high. It is necessary to combine artificial intelligence algorithms to further promote the implementation of technologies such as device status prediction of digital twins, simulation and deduction of power grid twin scenarios, and control and feedback of twin scenarios and physical scenarios. In terms of distribution network applications, there is currently a lack of composite embedded mapping and linkage of power semantics such as equipment electrical topology, a lack of numerical analysis and prediction of internal characteristics and evolution laws, and it is difficult to grow synchronously with the dynamically evolving power system.

[0008] In addition, the digital twin model still has the problem of insufficient application demonstration value. At present, a large number of digital twin pilot applications focus on the display of three-dimensional display scenes and data superposition scenes, which do not reflect the core capabilities of digital twin multi-source data-driven simulation and deduction, fail to focus on solving the pain points of the development of new power systems, have not yet been truly integrated into the production process, are not practical enough, and have insufficient application value. The digital twin application of distribution network equipment management is mainly concentrated in scene display, status monitoring, and real-time data applications based on threshold alarms. The perception and interpretation of key indicator characteristics and the simulation and deduction capabilities after real-time data analysis are still insufficient. Functions such as intelligent judgment and active early warning have not yet been put into practical use, so it is necessary to further expand the core application scenarios.

[0009] In response to the above problems, there is an urgent need for a method and system for assessing the risk of equipment cascading failures based on the distribution network twin model. Summary of the invention

[0010] In order to solve the deficiencies in the prior art, the present invention provides a method and system for assessing the risk of equipment cascading failures based on a distribution network twin model. The method simulates and analyzes the risks of each stage of cascading failures in a digital twin distribution network, obtains the low voltage risk and overload risk at each stage of the cascading failure process, and finally obtains the total risk value of cascading failures under various operating conditions.

[0011] The present invention adopts the following technical solution.

[0012] The present invention relates to a method for assessing the risk of equipment cascading failures based on a distribution network twin model. The method comprises the following steps: generating N-1-k expected accident sets and N-1-1 expected accident sets for different initial expected shutdown components when they fail, respectively, according to the failure state of each equipment component and the topological relationship between equipment components in the distribution network; wherein, the distribution network topology is extracted to determine the topological connection mode of the initial expected shutdown components in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode; the failure probability of each equipment component in the distribution network is calculated, and for each expected accident set, the low voltage risk of all nodes in the expected accident set and the overload risk of all lines in the expected accident set are calculated to generate the distribution network cascading failure risk for each expected accident set; and the importance of different expected accident sets is sorted based on the distribution network cascading failure risk of each expected accident set.

[0013] Preferably, N-1-k anticipated accident sets and NI-1 anticipated accident sets are generated for different initial anticipated shutdown components when they fail, based on the failure state of each equipment component and the topological relationship between equipment components in the distribution network, including: enumerating the initial anticipated shutdown components in the distribution network, and extracting the topological connection mode of the initial anticipated shutdown components in the distribution network according to a pre-generated distribution network twin model; judging the type of anticipated accidents that may be caused by the initial anticipated shutdown components when they are shut down based on the topological connection mode; generating anticipated accidents for the initial anticipated shutdown components based on the anticipated accident types, and extracting a set of all equipment components that may fail for the current anticipated accident to form an anticipated accident set.

[0014] Preferably, the distribution network topology is extracted to determine the topological connection mode of the initial expected shutdown element in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode, including: querying in the topological connection mode, if the initial expected shutdown element is connected to multiple transmission lines installed on the same tower, then an N-1-k expected accident set is generated for the initial expected shutdown element, k is the number of transmission lines; if the initial expected shutdown element is linked to multiple cables laid in the same channel, then an N-1-k expected accident set is generated for the initial expected shutdown element, k is the number of cables; otherwise, an N-1-1 expected accident set is generated for the weak links in the distribution network.

[0015] Preferably, the distribution network topology is extracted to determine the topological connection mode of the initial expected shutdown element in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode, including: in the distribution network twin model, querying the connection equipment on the other side of the multi-circuit transmission line corresponding to the initial expected shutdown element, and adding the initial expected shutdown element and the connection equipment on the other side to the N-1-k expected accident set; in the distribution network twin model, querying the connection equipment on the other side of the multi-circuit cable corresponding to the initial expected shutdown element, and adding the initial expected shutdown element and the connection equipment on the other side to the N-1-k expected accident set; in the distribution network twin model, starting from the shutdown of the initial expected shutdown element, the distribution network flow is calculated to obtain the weak link of the distribution network, and all equipment elements corresponding to the weak link of the distribution network are extracted to generate an N-1-1 expected accident set.

[0016] Preferably, calculating the failure probability of each device component in the distribution network includes: collecting the historical failure rate, device status evaluation, and real-time meteorological information of each device component in the distribution network; establishing a calculation model, and inputting the collected information into the calculation model to calculate the real-time failure rate of each device component, thereby generating the failure probability Q for different device components n under different scenarios n .

[0017] Preferably, for each anticipated accident set, the low voltage risk of all nodes in the anticipated accident set and the overload risk of all lines are calculated to generate the distribution network cascading failure risk for each anticipated accident set, including: using the failure probability of each equipment component to simulate the system state probability of failure of the initial anticipated shutdown component; using the system state probability and system load state to calculate the distribution network flow, and based on the distribution network flow, obtain the low voltage risk of all nodes in the current system state and the overload risk of all lines.

[0018] Preferably, the failure probability of each equipment component is used to simulate the system state probability of failure of the initially expected shutdown component, including: the system state probability is:

[0019]

[0020] In the formula, E i is the system state when the i-th equipment component fails as the initial expected outage component;

[0021] n is the equipment component number,

[0022] N f is the set of all failed components in the current system state,

[0023] N is the set of all components in the current system state,

[0024] Q n is the failure probability of the nth equipment component, P n is the working probability of the nth device component.

[0025] Preferably, the distribution network flow is calculated using the system state probability and the system load state, and the low voltage risk of all nodes and the overload risk of all lines under the current system state are obtained based on the distribution network flow, including: the low voltage risk of the jth node when the i-th accident occurs is:

[0026]

[0027] Where P t (V ij |E i , L) is the system state E i When the system load state is L, the voltage of the jth node is V ij The probability of

[0028] Sev(V ij ) is the voltage of the jth node V ij Fault severity index at the time of

[0029] The low voltage risk of the mth line when the i-th accident occurs is:

[0030]

[0031] Where P t (S im |E i , L) is the system state E i When the system load state is L, the power flow of the mth line is S im The probability of

[0032] Sev(S im ) is the flow of the mth line, S im Fault severity indicator at the time.

[0033] Preferably, the low voltage risk of all nodes and the overload risk of all lines in the expected accident set are calculated to generate the distribution network cascading failure risk for each expected accident set, including: the distribution network cascading failure risk under the current expected accident set E is:

[0034]

[0035] In the formula, It is the set of all nodes and all lines in the distribution network covered by the preset cascading failure level.

[0036] The second aspect of the present invention relates to a risk assessment system for equipment cascading failures based on a distribution network twin model using the method in the first aspect of the present invention; the system includes a generation module, a calculation module and a sorting module; wherein the generation module is used to generate N-1-k expected accident sets and NI-1 expected accident sets for different initial expected shutdown components when they fail, according to the failure state of each equipment component and the topological relationship between equipment components in the distribution network; wherein the distribution network topology is extracted to determine the topological connection mode of the initial expected shutdown components in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode; the calculation module is used to calculate the failure probability of each equipment component in the distribution network, and for each expected accident set, the low voltage risk of all nodes in the expected accident set and the overload risk of all lines in the expected accident set are calculated to generate the distribution network cascading failure risk for each expected accident set; the sorting module is used to sort the importance of different expected accident sets based on the distribution network cascading failure risk of each expected accident set.

[0037] A third aspect of the present invention relates to a terminal, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method in the first aspect of the present invention.

[0038] A fourth aspect of the present invention relates to a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in the first aspect of the present invention.

[0039] The beneficial effect of the present invention is that, compared with the prior art, the risk assessment method and system of equipment cascading failure based on the distribution network twin model in the present invention obtains the low voltage risk and overload risk at each stage of the cascading failure process through simulation and risk analysis of each stage of the digital twin distribution network cascading failure, and finally obtains the total risk value of cascading failure under various operating conditions. The present invention takes into account the fault conditions of the distribution network transmission lines erected on the same tower and the transmission cables laid in the same trench, generates the initial fault set using the dynamic N-1 state selection method, and finally sorts the risk values ​​of various states to compare their importance.

[0040] The beneficial effects of the present invention also include:

[0041] 1. The method of the present invention solves the problems of "where is the power outage, where is the load, where is the risk" in the distribution network, and improves the level of refined management services. By building a digital twin model of distribution network equipment, it can achieve panoramic visualization of distribution equipment operation, full equipment status perception, automatic fault diagnosis and prediction, and auxiliary decision-making for operation and maintenance.

[0042] 2. The present invention ensures that the digital twin model of the power grid is based on the characteristics of digital twins, namely, "strong real-time, high fidelity, and self-evolution". Starting from actual business needs, it realizes the multi-dimensional fusion application of digital twins based on the accumulated results of the three-dimensional data model of the power grid, the business data of power grid resources, and typical business scenarios of power grid equipment management, and further shortens the perception, analysis, application, and interaction links of power grid data.

[0043] 3. In terms of distribution networks, the present invention clarifies how to construct the digital twin system, breaks through the digital twin model aggregation mapping technology, supports the linkage analysis between equipment, topologies, and systems in multiple scenarios and multiple links of the "line-transformer" relationship, and provides a technical basis for building grid status perception, real-time control, and optimized collaborative applications for large-scale digital twin aggregation.

[0044] 4. In view of the problem of N-1 fault risk in digital twin distribution network, the present invention studies the impact of a certain device failure in the distribution network on other devices. Based on the collected and processed distribution network risk data, the severity of plant and station shutdown, busbar pressure loss, power supply to important users, load loss, and flow over-limit are calculated through static safety analysis and dynamic safety analysis, and combined with the risk rule base, the risk grading results of digital twin distribution network are obtained, so as to conduct chain failure reasoning analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of a method for obtaining an expected accident set in a method for risk assessment of equipment cascading failures based on a distribution network twin model of the present invention;

[0046] Figure 2It is a schematic diagram of fault removal in a method for risk assessment of equipment cascading failures based on a distribution network twin model of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present invention clearer and more correct, the technical solution of the present invention is described in detail through multiple specific implementations below. It is worth noting that the embodiments adopted in the present invention are only used to explain the present invention and are not used to limit the content of the present invention.

[0048] The present invention combines the twin linkage reasoning of distribution network equipment and topological network to analyze the risk of distribution network chain failure, and proposes an overall route for the equipment chain failure risk assessment method based on the distribution network twin model. In the digital twin scenario, the present invention first proposes a chain failure reasoning analysis method based on dynamic N-1 risk and an equipment chain failure risk assessment method based on multi-source data to enhance the fault diagnosis capability of the digital twin. Then, a distribution network chain failure prediction and deduction analysis technology is proposed, so that fault prevention can be carried out through accurate inference of the digital twin. Finally, the above technologies are combined to form a set of twin chain failure reasoning and analysis methods that combine distribution network equipment and topological networks, thereby realizing fault diagnosis, prediction and deduction of distribution networks.

[0049] Figure 1 Schematic diagram of the method for obtaining the expected accident set in the equipment cascading failure risk assessment method based on the distribution network twin model of the present invention. Figure 1 As shown, the first aspect of the present invention relates to a method for assessing the risk of equipment cascading failures based on a distribution network twin model, and the method includes steps 1 to 3.

[0050] Step 1: Generate N-1-k expected accident sets and N-1-1 expected accident sets for different initial expected shutdown components failures according to the failure state of each device component and the topological relationship between the device components in the distribution network. The topology of the distribution network is extracted to determine the topological connection mode of the initial expected shutdown components in the distribution network, and the N-1-k expected accident set or the N-1-1 expected accident set is generated based on the topological connection mode.

[0051] Dynamic N-1 risk is defined as a situation where, after a device in the distribution network stops operating for some reason, another device or multiple devices with a close electrical distance to it fails, which may cause problems such as equipment overload, load loss, and loss of stability of related network segments or even the entire distribution network. It is a combination of the probability of occurrence and the severity of the consequences.

[0052] The dynamic N-1 state selection method first enumerates the states according to the system state probability, and then dynamically generates the N-1-1 and N-1-k anticipated accident sets based on the actual distribution network operation conditions.

[0053] Preferably, N-1-k anticipated accident sets and NI-1 anticipated accident sets are generated for different initial anticipated shutdown components when they fail, based on the failure state of each equipment component and the topological relationship between equipment components in the distribution network, including: enumerating the initial anticipated shutdown components in the distribution network, and extracting the topological connection mode of the initial anticipated shutdown components in the distribution network according to a pre-generated distribution network twin model; judging the type of anticipated accidents that may be caused by the initial anticipated shutdown components when they are shut down based on the topological connection mode; generating anticipated accidents for the initial anticipated shutdown components based on the anticipated accident types, and extracting a set of all equipment components that may fail for the current anticipated accident to form an anticipated accident set.

[0054] Randomly select a component in the distribution network as the initial expected outage component to generate an expected accident. Traverse all distribution network equipment components to obtain all possible expected accidents in the distribution network.

[0055] The distribution network twin model is constructed according to the actual state of the distribution network, extracts real-time information from the distribution network equipment, obtains the real-time operating state of the distribution network through simulation, and deduces the distribution network equipment failures on this basis.

[0056] Preferably, the distribution network topology is extracted to determine the topological connection mode of the initial expected shutdown element in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode, including: querying in the topological connection mode, if the initial expected shutdown element is connected to multiple transmission lines installed on the same tower, then an N-1-k expected accident set is generated for the initial expected shutdown element, k is the number of transmission lines; if the initial expected shutdown element is linked to multiple cables laid in the same channel, then an N-1-k expected accident set is generated for the initial expected shutdown element, k is the number of cables; otherwise, an N-1-1 expected accident set is generated for the weak links of the distribution network.

[0057] After enumerating the initial N-1 state according to the probability of the distribution network state, the topological analysis is used to determine whether there are multiple transmission lines erected on the same tower or multiple cables laid in the same trench around the initial expected outage components. If so, an N-1-k expected accident set is generated; if not, an N-1-1 expected accident set is generated. Through distribution network vulnerability analysis and key equipment identification, the weak links of the distribution network after N-1 are obtained, and the weak equipment fault set is generated. Dynamic N-1 analysis is performed through high-correlation fault sets and weak equipment fault sets with similar electrical distances.

[0058] In the above method, the transmission lines and cables connected to the initial expected shutdown element will be directly impacted after being affected by the state of the expected shutdown element, thus causing the failure or state of the equipment on the other side of the transmission line or cable to exceed the limit. Through the topological connection structure, predicting the failure of the equipment on the other side of the transmission line can more accurately and quickly obtain possible risks, ensuring real-time risk analysis and rapid fault cutting. Therefore, the expected fault set obtained by the method of the present invention comprehensively analyzes the source of the fault, overcoming the problem that it is difficult to find the source of the fault and it is difficult to accurately implement fault removal after the distribution network chain fault occurs.

[0059] Preferably, the distribution network topology is extracted to determine the topological connection mode of the initial expected shutdown element in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode, including: in the distribution network twin model, querying the connection equipment on the other side of the multi-circuit transmission line corresponding to the initial expected shutdown element, and adding the initial expected shutdown element and the connection equipment on the other side to the N-1-k expected accident set; in the distribution network twin model, querying the connection equipment on the other side of the multi-circuit cable corresponding to the initial expected shutdown element, and adding the initial expected shutdown element and the connection equipment on the other side to the N-1-k expected accident set; in the distribution network twin model, starting from the shutdown of the initial expected shutdown element, the distribution network flow is calculated to obtain the weak link of the distribution network, and all equipment elements corresponding to the weak link of the distribution network are extracted to generate an N-1-1 expected accident set.

[0060] When obtaining the N-1-k expected accident set, the probability of equipment failure can be used for sorting. For a certain expected outage component or a certain system state, the failure probability of other interlocking equipment is analyzed, and the equipment with a higher failure probability is sorted first, and the equipment with a lower failure probability is sorted last. The system state is selected through quick sorting to generate the corresponding fault set. In addition, if the probability of failure of the equipment on the other side of the multi-circuit transmission lines and multi-circuit cables is small during sorting, these equipment are still added to the expected accident set.

[0061] The probability of equipment failure, that is, the failure probability of a certain equipment component, is calculated in advance according to certain rules. In the embodiment of the present invention, the probability of equipment failure is also used to calculate the probability and degree of occurrence of distribution network cascading failure risk.

[0062] Step 2: Calculate the failure probability of each device component in the distribution network, and for each expected accident set, calculate the low voltage risk of all nodes and the overload risk of all lines in the expected accident set to generate the distribution network cascading failure risk for each expected accident set.

[0063] Preferably, calculating the failure probability of each device component in the distribution network includes: collecting the historical failure rate, device status evaluation, and real-time meteorological information of each device component in the distribution network; establishing a calculation model, and inputting the collected information into the calculation model to calculate the real-time failure rate of each device component, thereby generating the failure probability Q for different device components n under different scenarios n .

[0064] Equipment status evaluation can be an evaluation quantity generated by power grid dispatchers, operation and maintenance personnel based on actual conditions, or it can be an indirect quantity calculated based on historical data related to the distribution network, used to evaluate the safety and reliability of equipment or equipment failure. Real-time meteorological information is obtained based on the geographical location of equipment components. By collecting weather forecast information, the performance of equipment under similar climates can be obtained. The historical failure rate of equipment components is obtained by analyzing the historical failure data of the equipment.

[0065] The present invention supports the establishment of data models. It is conceivable that the above historical data is trained by a common deep learning model and the like to obtain model parameters. The trained model supports the input of real-time data and outputs the failure probability of each device component, that is, the failure probability. Different scenarios may include different weather conditions, different load conditions, etc.

[0066] Preferably, for each anticipated accident set, the low voltage risk of all nodes in the anticipated accident set and the overload risk of all lines are calculated to generate the distribution network cascading failure risk for each anticipated accident set, including: using the failure probability of each equipment component to simulate the system state probability of failure of the initial anticipated shutdown component; using the system state probability and system load state to calculate the distribution network flow, and based on the distribution network flow, obtain the low voltage risk of all nodes in the current system state and the overload risk of all lines.

[0067] Each anticipated accident set in the present invention includes a plurality of equipment components and fault information corresponding to the components.

[0068] After the dynamic N-1 fault reasoning analysis of the digital twin distribution network, it is also necessary to study the risk probability and risk level of the cascading failure in order to implement certain countermeasures for specific risks. Therefore, the cascading failure risk assessment of distribution network equipment is carried out. The risk of the distribution network is a comprehensive measure of probability and severity. A cascading failure is a collection of a series of successive failures. The risk of a cascading failure in the power grid should consider the probability of occurrence of various fault conditions and their corresponding consequences. The cascading failure process violates safety criteria such as current and voltage. Therefore, the cascading failure risk index is the accumulation of low voltage risks and line overload risks at various stages of the cascading failure development.

[0069] In the present invention, the cascading failure risk index is evaluated by the low voltage risk of all nodes in the cascading failure search range and the overload risk of all lines in the system state.

[0070] Preferably, the failure probability of each equipment component is used to simulate the system state probability of failure of the initially expected shutdown component, including: the system state probability is:

[0071]

[0072] In the formula, E i is the system state when the i-th equipment component fails as the initial expected outage component;

[0073] n is the equipment component number,

[0074] N f is the set of all failed components in the current system state,

[0075] N is the set of all components in the current system state,

[0076] Q n is the failure probability of the nth equipment component, P n is the working probability of the nth device component.

[0077] In normal state, all components are running, N f =0, then the system state probability is:

[0078]

[0079] As mentioned above, all system states are obtained by enumerating different initial expected shutdown components. In one embodiment, it is supported to slightly adjust the failure probability of each component inferred by the data model, so as to ensure that the failure probability of each model caused by different reasons has different values. Thus, when all system states are enumerated, the values ​​of each system state probability calculated are mutually exclusive, and the cumulative state probability of the system is the sum of all state probabilities, which is:

[0080]

[0081] Where G is the set of all failure states.

[0082] Preferably, the distribution network flow is calculated using the system state probability and the system load state, and the low voltage risk of all nodes and the overload risk of all lines under the current system state are obtained based on the distribution network flow, including: the low voltage risk of the jth node when the i-th accident occurs is:

[0083]

[0084] Where Pt (V ij |E i , L) is the system state E i When the system load state is L, the voltage of the jth node is V ij The probability of

[0085] Sev(V ij ) is the voltage of the jth node V ij Fault severity index at the time of

[0086] The low voltage risk of the mth line when the i-th accident occurs is:

[0087]

[0088] Where P t (S im |E i , L) is the system state E i When the system load state is L, the power flow of the mth line is S im The probability of

[0089] Sev(S im ) is the flow of the mth line, S im Fault severity indicator at the time.

[0090] As described in the formula, the low voltage risk of the digital twin distribution network reflects the possibility and consequences of low voltage at the node due to a fault in the distribution network. The transmission line overload risk of the digital twin distribution network reflects the possibility and consequences of line overload caused by load redistribution after the fault occurs in the distribution network. Among them, according to the system operation state and the system load state, the possible voltage value of each node in the regional range is obtained by flow calculation. Because under a certain expected accident set, although the initial failed equipment components are determined, different system states such as load conditions, weather conditions, and subtle differences in equipment operation status may cause chain failures to develop in different directions. Therefore, the fault evolution results are diverse, and each node may have a variety of different voltage states. For a certain voltage state, there is a probability that this voltage state exists after the accident occurs. This voltage state will cause the equipment on the node to be at different risks. According to the characteristics of the equipment on the node itself, the equipment may be in normal operation, different levels of equipment over-limit, serious failure, etc. at a certain voltage. The present invention uses severity indicators to quantify different situations, so as to obtain the node voltage risk value by calculation. In a similar way, the system flow is calculated to obtain the overload risk of multiple transmission lines.

[0091] As described in the above formula, if there are multiple different deduction results in the actual calculation, the risk of each deduction result is summed up again to obtain the total risk of all different possible voltage states of a node, and the total risk of all different possible flow states of a line.

[0092] Preferably, the low voltage risk of all nodes and the overload risk of all lines in the expected accident set are calculated to generate the distribution network cascading failure risk for each expected accident set, including:

[0093] The distribution network cascading failure risk under the current expected accident set E is:

[0094]

[0095] In the formula, It is the set of all nodes and all lines in the distribution network covered by the preset cascading failure level.

[0096] The risk of cascading failures in the digital twin distribution network reflects the probability of a system failure and the consequences of the cascading failure process. It is the accumulation of risk values ​​at each stage of the chain reaction process. According to the different levels of cascading failures, the area and range that the chain reaction process can affect are obtained, and the risks of all nodes and lines in the area are calculated and summed.

[0097] Through the simulation and risk analysis of each stage of the cascading failure of the digital twin distribution network, the low voltage risk and overload risk at each stage of the cascading failure process are obtained, and finally the total risk value of the cascading failure under various operating conditions is obtained.

[0098] Step 3: Based on the distribution network cascading failure risk of each expected accident set, the importance of different expected accident sets is ranked.

[0099] Considering the fault conditions of the distribution network transmission lines erected on the same tower and the transmission cables laid in the same trench, the dynamic N-1 state selection method is used to generate the initial fault set, and finally the risk values ​​of various states are ranked and compared to compare their importance. Thus, the importance of each different fault can be deduced through the digital twin model, so as to pre-redundant backup or risk avoidance of more important equipment that is more likely to cause risks.

[0100] In addition, the digital twin model can be used to randomly trigger events from the fault set of the present invention and set chain reaction conditions. For different chain reaction levels, the power flow calculation method is used to determine whether the power grid model can automatically converge, and the low voltage risk of different nodes and the load risk of different lines, as well as the chain failure risk value in the area, are calculated at the same time.

[0101] Figure 2Schematic diagram of fault removal in a method for risk assessment of equipment cascading failures based on a distribution network twin model of the present invention. Figure 2 As shown in the figure, if the bus where the node with the largest low voltage risk value is located is a load bus and the risk value is greater than the threshold, the load on the corresponding bus is cut off. If the bus where the node with the largest low voltage risk is located is a generator bus, the current generator set is cut off. In addition, the method also supports setting different thresholds for nodes at different locations. If the set threshold is exceeded, the corresponding loads on multiple nodes can be cut off. In addition, the amount of load cut can be set according to the amount exceeding the set threshold, or the method of iterative cut can be adopted to ensure that the load is cut off in small amounts until all loads are cut off.

[0102] The line with the maximum overload risk value is also cut off in the same way until all lines are cut off.

[0103] When calculating the power flow and determining whether the distribution network meets the convergence conditions, if no fault occurs in a certain state, the calculated cascading failure risk value in this state is set to the maximum value as the threshold reference in the dynamic dispatching process of the power grid.

[0104] The second aspect of the present invention relates to a risk assessment system for equipment cascading failures based on a distribution network twin model using the method in the first aspect of the present invention; the system includes a generation module, a calculation module and a sorting module; wherein the generation module is used to generate N-1-k expected accident sets and NI-1 expected accident sets for different initial expected shutdown components when they fail, according to the failure state of each equipment component and the topological relationship between equipment components in the distribution network; wherein the distribution network topology is extracted to determine the topological connection mode of the initial expected shutdown components in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode; the calculation module is used to calculate the failure probability of each equipment component in the distribution network, and for each expected accident set, the low voltage risk of all nodes in the expected accident set and the overload risk of all lines in the expected accident set are calculated to generate a distribution network cascading failure risk for each expected accident set; the sorting module is used to sort the importance of different expected accident sets based on the distribution network cascading failure risk of each expected accident set.

[0105] A second aspect of the present invention relates to a terminal, comprising a processor and a storage medium; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the steps of the method in the first aspect of the present invention.

[0106] A third aspect of the present invention relates to a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in the first aspect of the present invention.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention is described in detail with reference to the above embodiments, it should be understood by those skilled in the art that the technical solutions of the present invention still include contents that can modify or replace the specific embodiments of the present invention. Any modification or replacement that does not depart from the spirit and scope of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A method for assessing the risk of equipment cascading failures based on a distribution network twin model, characterized in that: The method comprises the following steps: According to the failure state of each equipment component and the topological relationship between equipment components in the distribution network, the N-1-k expected accident set and the NI-1 expected accident set are generated for different initial expected shutdown component failures; The distribution network topology is extracted to determine the topological connection mode of the initial expected outage components in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode; Calculate the failure probability of each device component in the distribution network, and for each expected accident set, calculate the low voltage risk of all nodes and the overload risk of all lines in the expected accident set to generate the distribution network cascading failure risk for each expected accident set; Based on the distribution network cascading failure risk of each anticipated accident set, the importance of different anticipated accident sets is ranked.

2. According to claim 1, a method for assessing the risk of equipment cascading failures based on a distribution network twin model is characterized in that: The N-1-k anticipated accident set and the NI-1 anticipated accident set for different initial anticipated shutdown component failures are generated according to the failure state of each device component and the topological relationship between the device components in the distribution network, including: Enumerate the initial expected outage components in the distribution network, and extract the topological connection mode of the initial expected outage components in the distribution network according to the pre-generated distribution network twin model; Determining the type of anticipated accident that may be caused by the initial anticipated outage component when it is out of service based on the topological connection mode; Based on the anticipated accident type, an anticipated accident for the initial anticipated outage component is generated, and for the current anticipated accident, a set of all equipment components that may fail is extracted to form an anticipated accident set.

3. According to claim 2, a method for assessing the risk of equipment cascading failures based on a distribution network twin model is characterized in that: The extracting of the distribution network topology determines the topological connection mode of the initial expected outage components in the distribution network, and generating an N-1-k expected accident set or an N-1-1 expected accident set based on the topological connection mode, including: Inquiring in the topological connection mode, if the initial expected outage component is connected to multiple transmission lines erected on the same tower, then generating an N-1-k expected accident set for the initial expected outage component, where k is the number of transmission lines; If the initial expected outage component is connected to multiple cables laid in the same channel, then an N-1-k expected accident set for the initial expected outage component is generated; k is the number of cables; Otherwise, an N-1-1 anticipated accident set is generated for the weak links of the distribution network.

4. According to claim 3, a method for assessing the risk of equipment cascading failures based on a distribution network twin model is characterized in that: The extracting of the distribution network topology determines the topological connection mode of the initial expected outage components in the distribution network, and generating an N-1-k expected accident set or an N-1-1 expected accident set based on the topological connection mode, including: In the distribution network twin model, query the connection equipment on the other side of the multi-circuit transmission line corresponding to the initial expected outage element, and add the initial expected outage element and the connection equipment on the other side to the N-1-k expected accident set; In the distribution network twin model, query the connection device on the other side of the multi-circuit cable corresponding to the initial expected outage component, and add the initial expected outage component and the connection device on the other side to the N-1-k expected accident set; In the distribution network twin model, starting from the outage of the initial expected outage component, the distribution network flow is calculated to obtain the weak link of the distribution network, and all equipment components corresponding to the weak link of the distribution network are extracted to generate an N-1-1 expected accident set.

5. According to claim 4, a method for assessing the risk of equipment cascading failures based on a distribution network twin model is characterized in that: The calculation of the failure probability of each device component in the distribution network includes: Collect historical failure rates, equipment status evaluation, and real-time meteorological information for each equipment component in the distribution network; Establish a calculation model and input the collected information into the calculation model to calculate the real-time failure rate of each device component, thereby generating the failure probability Q for different device components n under different scenarios n .

6. According to claim 5, a method for assessing the risk of equipment cascading failures based on a distribution network twin model is characterized in that: For each expected accident set, the low voltage risk of all nodes and the overload risk of all lines in the expected accident set are calculated to generate the distribution network cascading failure risk for each expected accident set, including: Using the failure probability of each equipment component, simulate the system state probability of failure of the initial expected shutdown component; The distribution network flow is calculated using the system state probability and the system load state, and based on the distribution network flow, the low voltage risk of all nodes and the overload risk of all lines in the current system state are obtained.

7. The method for assessing the risk of equipment cascading failures based on the distribution network twin model according to claim 6 is characterized in that: The method of using the failure probability of each equipment component to simulate the system state probability of failure of the initially expected shutdown component includes: The system state probability is: In the formula, E i is the system state when the i-th equipment component fails as the initial expected outage component; n is the equipment component number, N f is the set of all failed components in the current system state, N is the set of all components in the current system state, Q n is the failure probability of the nth equipment component, P n is the working probability of the nth device component.

8. The method for assessing the risk of equipment cascading failures based on the distribution network twin model according to claim 7 is characterized in that: The method of calculating the distribution network flow by using the system state probability and the system load state, and obtaining the low voltage risk of all nodes and the overload risk of all lines under the current system state based on the distribution network flow, includes: The low voltage risk of the jth node when the i-th accident occurs is: Where P t (V ij |E i ,L) is the system state E i When the system load state is L, the voltage of the jth node is V ij The probability of Sev(V ij ) is the voltage of the jth node V ij Fault severity index at the time of The low voltage risk of the mth line when the i-th accident occurs is: Where P t (S im |E i ,L) is the system state E i When the system load state is L, the power flow of the mth line is S im The probability of Sev(S im ) is the flow of the mth line, S im Fault severity indicator at the time.

9. The method for risk assessment of equipment cascading failures based on a distribution network twin model according to claim 8 is characterized in that: The calculation of the low voltage risk of all nodes and the overload risk of all lines in the expected accident set to generate the distribution network cascading failure risk for each expected accident set includes: The distribution network cascading failure risk under the current expected accident set E is: Where K is the set of all nodes and all lines in the distribution network covered under the preset cascading failure level.

10. A device cascading failure risk assessment system based on a distribution network twin model using the method described in any one of claims 1 to 9; characterized in that: The system includes a generation module, a calculation module and a sorting module; wherein, The generation module is used to generate N-1-k expected accident sets and NI-1 expected accident sets for failure of different initial expected shutdown components according to the failure state of each equipment component and the topological relationship between equipment components in the distribution network; The distribution network topology is extracted to determine the topological connection mode of the initial expected outage components in the distribution network, and an N-1-k expected accident set or an N-1-1 expected accident set is generated based on the topological connection mode; The calculation module is used to calculate the failure probability of each device component in the distribution network, and for each expected accident set, calculate the low voltage risk of all nodes in the expected accident set and the overload risk of all lines, so as to generate the distribution network cascading failure risk for each expected accident set; The ranking module is used to rank the importance of different anticipated accident sets based on the distribution network cascading failure risk of each anticipated accident set.

11. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1-9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

Images