A method and system for emergency operation and maintenance of power system based on parallel DC power supply
Through the emergency operation and maintenance method based on parallel DC power supply, the network topology diagram of the power system is monitored in real time, and the internal network and topology diagram of the power supply are reconstructed, which solves the problem of slow reconstruction of power system faults, and achieves fast response and high efficiency Operation and maintenance.
Patent Information
- Application Number
- CN202411156151.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-22
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-22
AI Technical Summary
The existing power system is slow to reconstruct after a failure, resulting in a long system interruption time, and it is difficult to predict the fault in advance, and it is impossible to quickly adjust the network topology to avoid failure.
Using an emergency operation and maintenance method based on parallel DC power supply, by constructing parallel DC power supply and network topology diagrams, monitoring node parameters in real time, determining abnormal status, and performing internal network reconstruction of power supply and local or global reconstruction of topology diagrams to quickly respond to faults .
It realizes the prevention of hidden dangers in advance before the failure occurs, responding to sudden failures quickly, reducing system interruption time, and improving the stability and operation and maintenance efficiency of the power system.
Smart Images

Figure CN118693777B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of power system operation and maintenance, and in particular to a power system emergency operation and maintenance method and system based on a parallel type direct current power supply. Background Art
[0002] At present, with the substantial growth in electricity demand, the power system network has become increasingly complex and the possibility of failure has increased dramatically, posing greater challenges to emergency operation and maintenance requirements.
[0003] Power network reconstruction means disconnecting the faulty line after a line fails, changing the opening and closing states of switches in the line, and optimizing the structure of the distribution network to ensure that as many loads as possible can be restored as quickly as possible.
[0004] In the prior art, when a fault occurs in the operation and maintenance of the power system, the power supply is cut off or the connection between the faulty device and other devices is cut off, and then the entire power network is reconstructed. However, this operation and maintenance method is to remedy the fault after it occurs, and the reconstruction speed is slow, which is prone to excessive system interruption time. The operation and maintenance process also has high requirements for the stability of the system's power supply.
[0005] However, the process of load failure is often not sudden. When its operating state drops to the alarm threshold, the process is continuous in time dimension. Therefore, how to predict the possibility of power system failure in advance, quickly adjust the system network topology to avoid the occurrence of failure, and how to respond to emergency operation and maintenance of sudden failures are difficult problems that need to be solved urgently. Summary of the invention
[0006] In view of the problems existing in the above-mentioned prior art, the present invention provides an emergency operation and maintenance method and system for an electric power system based on a parallel DC power supply, the purpose of which is to capture the abnormal trend of the electric power system, optimize the power network topology in advance to avoid the occurrence of faults, and realize timely and rapid emergency operation and maintenance in the event of sudden faults.
[0007] To achieve the above object, the present invention provides the following technical solutions:
[0008] In a first aspect, the present invention provides an emergency operation and maintenance method for a power system based on a parallel DC power supply, comprising:
[0009] Step 1: Determine the number of batteries required to build a parallel DC power supply based on the load demand and operation and maintenance current of the power system;
[0010] Step 2, constructing a network topology diagram corresponding to the power system according to the connection relationship between the power supply end and each load end in the power system, collecting real-time operating parameters of each node in the network topology diagram and determining abnormal nodes;
[0011] Step 3, when the abnormal node is a power supply end, changing the connection mode of the single battery in the parallel DC power supply to reconstruct the internal network of the power supply;
[0012] When the abnormal node is a load end, determining the abnormal state type of the power system based on the real-time operating parameters, the abnormal state type includes a hidden danger state and a fault state;
[0013] When the abnormal state is determined to be a hidden danger state, determining the hidden danger impact range based on the real-time operating parameters of each node in the network topology map, and partially reconstructing the network topology map based on the hidden danger impact range;
[0014] When the abnormal state is determined to be a fault state, the abnormal node is removed and the network topology diagram is globally reconstructed to implement emergency operation and maintenance.
[0015] Preferably, the step 1 of determining the required number of batteries based on the load demand and operation and maintenance current of the power system to construct a parallel DC power supply includes:
[0016] Determining a first number of storage batteries required for the power system based on the rated total energy of the primary side of the power system and the rated capacity of a single storage battery;
[0017] determining a second number of storage batteries required for the power system based on a maximum load current when an abnormality occurs in the power system and a rated output current of a single storage battery;
[0018] The larger value between the first number and the second number is the number of storage batteries required to construct a parallel DC power supply.
[0019] Preferably, the step 3 of changing the connection mode of the single storage batteries in the parallel DC power supply to reconstruct the internal network of the power supply includes:
[0020] Disconnect the abnormal battery, and under the constraint that the total output voltage and current of the parallel DC power supply change minimally, change the connection mode and number of the remaining single batteries and the DC system bus to reconstruct the internal network of the power supply.
[0021] Preferably, the step 3 of determining the impact range of hidden dangers based on the real-time operating parameters of each node in the network topology diagram includes:
[0022] Obtaining the time point at which the hidden danger state of the power system occurs and the real-time operating parameters of the abnormal node at that time point, which are recorded as the hidden danger time point and the hidden danger parameter respectively;
[0023] The real-time operating parameters of the abnormal node within a preset time period before the hidden danger time point are collected piece by piece at a preset time step and recorded as a parameter sequence;
[0024] Based on the relative deviation between the parameter sequence and the hidden danger parameter, searching for the hidden danger trend starting point of the hidden danger state to determine the hidden danger trend duration;
[0025] In the network topology diagram, starting from the abnormal node according to the connection relationship between the nodes, the search for the first node within the impact range threshold is radially diffused outward;
[0026] The real-time operating parameters of the first node within the hidden danger trend duration and before the hidden danger trend starting point are respectively obtained, and the second node actually affected by the hidden danger state of the abnormal node is determined according to the difference value of the real-time operating parameters in the two time periods. The node set including all second nodes is the hidden danger impact range.
[0027] Preferably, searching for the starting point of the hidden danger trend of the hidden danger state based on the relative deviation between the parameter sequence and the hidden danger parameter to determine the hidden danger trend duration includes:
[0028] Respectively extracting the hidden danger parameter and the characteristic values of the real-time operation parameter in each time period in the parameter sequence;
[0029] Calculate the deviation value between the characteristic value of the real-time operating parameter and the characteristic value of the hidden danger parameter in each time period of the parameter sequence, and draw a deviation curve with time as the horizontal axis and the deviation value as the vertical axis;
[0030] The time point corresponding to the first inflection point of the deviation curve is the starting point of the hidden danger trend, and the time from the starting point of the hidden danger trend to the hidden danger time point is the duration of the hidden danger trend.
[0031] Preferably, the second node actually affected by the hidden danger state of the abnormal node is determined according to the difference value of the real-time operating parameters in two time periods, and the node set including all the second nodes is the hidden danger impact range, including:
[0032] For each of the first nodes,
[0033] The real-time operating parameters of the first node within the hidden danger trend duration are plotted into a trend state curve with time as the horizontal axis and the real-time operating parameters as the vertical axis; the real-time operating parameters of the first node before the starting point of the hidden danger trend are plotted into a normal state curve with time as the horizontal axis and the real-time operating parameters as the vertical axis;
[0034] A curve segment of the normal state curve having the same time length as the trend state curve and having the highest similarity is recorded as a comparison curve;
[0035] The comparison curve is equally divided according to the preset time step in the trend state curve; the similarity between the trend state curve and the comparison curve in each time step is calculated segment by segment in chronological order, and the number of affected segments is counted: if the similarity is less than a similarity threshold, the number of affected segments is increased by one;
[0036] The first node whose number of affected segments is greater than a preset threshold is taken as a second node, and a node set consisting of all second nodes is recorded as the hidden danger impact range.
[0037] Preferably, the reconstructing the network topology diagram specifically includes:
[0038] A multi-objective optimization reconstruction model is established with one or more of minimum network loss, maximum load balancing, maximum power supply capacity, and minimum number of switch operations as objective functions, and one or more of voltage constraints, power constraints, and objective function weight constraints as constraint conditions;
[0039] The binary particle swarm optimization algorithm is used to solve the optimal solution of the multi-objective optimization reconstruction model:
[0040] Different reconstruction schemes are abstracted into different particle positions; the dimension of the particle position is the number of switches in the network topology; the position of the particle in each dimension is 0 or 1, indicating the open or closed state of the corresponding switch in the network topology; the particle speed is defined as the probability of relative change of the particle position in the spatial dimension;
[0041] Based on the objective function and the constraint conditions, calculating the fitness of each particle in the current particle swarm;
[0042] Iteratively update the speed and position of particles to obtain the individual extreme value of each particle and the global extreme value of the particle group in each iteration process;
[0043] When the particle swarm converges or reaches the maximum number of iterations, the particle that currently obtains the global extreme value is output as the optimal solution of the multi-objective optimization reconstruction model, and the network topology graph is reconstructed according to the optimal solution.
[0044] Preferably, when the abnormal state is determined to be a hidden danger state, a multi-objective optimization reconstruction model under the hidden danger state is established, and a binary particle swarm optimization algorithm is used to locally reconstruct the network topology map based on the hidden danger impact range, specifically including:
[0045] Fixing the topological structure in the network topological graph outside the hidden danger impact range at the hidden danger time point, and randomly generating the topological structure in the network topological graph within the hidden danger impact range to form an initial particle swarm;
[0046] The velocity of a particle is defined as the probability of a change in the particle's position in each spatial dimension;
[0047] The probability of the switch state change is weighted based on the distance between each switch and the abnormal node, that is, the speed of the particle in each spatial dimension is weighted. The speed update formula is:
[0048] If the switch is within the range of the hidden danger,
[0049]
[0050] If the switch is outside the range of the hidden danger,
[0051]
[0052] in, Represents particles In the In the iteration The speed of the space, Represents particles In the In the iteration The speed of the space, Represents particles When the individual extreme value is obtained, The position of the dimensional space, The particle that achieves the global extreme value is The position of the dimensional space, Represents particles In the In the iteration The position of the dimensional space, and is the learning factor, and is a random number between 0 and 1;
[0053] is the inertia factor, which is negatively correlated with the number of iterations, indicating that as the number of iterations increases, the probability of particle position change decreases;
[0054] is the distance factor, , is negatively correlated with the distance between the switch and the abnormal node, indicating that the farther the distance between the switch and the abnormal node, the smaller the probability of the switch state changing.
[0055] Preferably, when the abnormal state is determined to be a fault state, a multi-objective optimization reconstruction model under the fault state is established, and the network topology graph is globally reconstructed using a binary particle swarm optimization algorithm, specifically including:
[0056] The abnormal nodes are removed from the network topology diagram, and an initial particle swarm is randomly generated. The speed of a particle is defined as the probability that the particle is at position 1 in each spatial dimension. The speed update formula is:
[0057]
[0058] in, Represents particles In the The fitness in the iteration; is the fitness dynamic factor, which is The fitness of the particle that obtains the global extreme value in the iteration The ratio of the average fitness of all particles in the iterations is negatively correlated; Represents particles In the In the iteration The speed of dimensional space; Represents particles In the In the iteration The speed of dimensional space; Represents particles When the individual extreme value is obtained, The position of the dimensional space; The particle that achieves the global extreme value is The position of the dimensional space; Represents particles In the In the iteration The position of the dimensional space; and is the learning factor; and A random number between 0 and 1.
[0059] In a second aspect, the present invention provides an emergency operation and maintenance system for a power system based on a parallel DC power supply, comprising:
[0060] A parallel DC power supply building module is used to build a parallel DC power supply based on the load demand of the power system and the number of batteries required for operation and maintenance current calculation;
[0061] A topology diagram building module is used to build a network topology diagram corresponding to the power system according to the connection relationship between the power supply end and each load end in the power system;
[0062] An abnormal node determination module is used to collect real-time operating parameters of each node in the network topology diagram to determine abnormal nodes;
[0063] The emergency operation and maintenance module is used to select a network reconstruction scheme according to the location of the abnormal node to perform emergency operation and maintenance of the power system, including:
[0064] When the abnormal node is a power supply end, changing the connection mode of the single battery in the parallel DC power supply to reconstruct the internal network of the power supply;
[0065] When the abnormal node is a load end, determining the abnormal state type of the power system based on the real-time operating parameters, the abnormal state type includes a hidden danger state and a fault state;
[0066] When the abnormal state is determined to be a hidden danger state, determining the hidden danger impact range based on the real-time operating parameters of each node in the network topology map, and partially reconstructing the network topology map based on the hidden danger impact range;
[0067] When the abnormal state is determined to be a fault state, the abnormal node is removed and the network topology diagram is globally reconstructed to implement emergency operation and maintenance.
[0068] The present invention provides a method and system for emergency operation and maintenance of a power system based on a parallel DC power supply, which has the following beneficial effects:
[0069] 1. The present invention designs the primary side of the power system by calculating the number of parallel power storage batteries to meet the emergency operation and maintenance requirements, ensuring that the system can continue to provide sufficient power in the early stage of abnormality. It replaces the conventional technology of fixed series-parallel connection of single batteries, avoiding the problem that when a single battery fails, the entire battery branch must be isolated for operation and maintenance, greatly improving the work efficiency and reliability of the system, and providing a good foundation for the realization of emergency operation and maintenance.
[0070] 2. In the emergency operation and maintenance process of the present invention, the abnormal state is first located to determine the terminal that needs to be reconstructed, including the reconstruction of the load line and the battery access adjustment of the parallel DC power supply. Then the abnormal state is analyzed, considering that different abnormal situations have different effects on the topological structure of the power system, different operation and maintenance methods are adopted for different degrees of abnormalities, so that the system can quickly adapt to different operating requirements in an emergency: when there is a hidden danger state, the hidden danger impact range is analyzed, and the topological structure within the hidden danger impact range is quickly adjusted without the need to reconstruct the entire system network, and the load of the hidden danger node is transferred to other lines. The possible abnormal state is handled in advance in a timely manner, avoiding the expansion of the abnormal range of the power system and causing substantial losses. At the same time, the cost of local reconstruction is low and the speed is high, which can better meet the emergency operation and maintenance needs; for sudden fault states, the present invention improves the reconstruction algorithm of the prior art, pays more attention to the real-time situation of the load during each reconstruction process, speeds up the reconstruction process, and minimizes the operation and maintenance time under the fault state. BRIEF DESCRIPTION OF THE DRAWINGS
[0071] Figure 1 It is a flow chart of an emergency operation and maintenance method of a power system based on a parallel type DC power supply of the present invention;
[0072] Figure 2 is a schematic diagram of the structure inside a parallel DC power supply in one embodiment of the present invention;
[0073] Figure 3 It is a structural block diagram of an emergency operation and maintenance system of an electric power system based on a parallel type DC power supply of the present invention. DETAILED DESCRIPTION
[0074] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0075] To facilitate understanding of this embodiment, first, a power system emergency operation and maintenance method based on a parallel DC power supply disclosed in an embodiment of the present invention is introduced in detail.
[0076] The embodiment of the present invention provides an emergency operation and maintenance method for a power system based on a parallel DC power supply, such as Figure 1 As shown, including:
[0077] Step 1: Determine the number of batteries required to build a parallel DC power supply based on the load demand and operation and maintenance current of the power system;
[0078] Step 2, constructing a network topology diagram corresponding to the power system according to the connection relationship between the power supply end and each load end in the power system, collecting real-time operating parameters of each node in the network topology diagram and determining abnormal nodes;
[0079] Specifically, the real-time operating parameters include but are not limited to the real-time voltage, current, and power of the power supply end and each load end.
[0080] Step 3, when the abnormal node is a power supply end, changing the connection mode of the single battery in the parallel DC power supply to reconstruct the internal network of the power supply;
[0081] When the abnormal node is a load end, determining the abnormal state type of the power system based on the real-time operating parameters, the abnormal state type includes a hidden danger state and a fault state;
[0082] In some embodiments, a hidden danger threshold and a fault threshold are set for the operating parameters to determine the abnormal state type of the power system, and when the real-time operating parameters reach the hidden danger / fault threshold, the power system issues a hidden danger / fault state alarm. In a more intuitive understanding, the hidden danger state is an abnormality with signs, and the fault state is an abnormality without signs.
[0083] When the abnormal state is determined to be a hidden danger state, determining the hidden danger impact range based on the real-time operating parameters of each node in the network topology map, and partially reconstructing the network topology map based on the hidden danger impact range;
[0084] When the abnormal state is determined to be a fault state, the abnormal node is removed and the network topology diagram is globally reconstructed to implement emergency operation and maintenance.
[0085] The embodiment of the present invention first utilizes the advantages of high safety and easy operation and maintenance of parallel batteries to construct a parallel DC power supply, which ensures the basic power supply performance of the power system from the root of the power system and reduces the cost and difficulty of later maintenance. When a hidden danger state occurs, since the impact range of the hidden danger state is small, there is no need to reconstruct the overall power system structure. The present invention focuses on optimizing the topological structure within the impact range of the hidden danger state to transfer the load of the abnormal node to other lines, and try to eliminate the hidden danger in advance at a lower cost and faster speed before the loss occurs, so as to avoid the hidden danger state from continuing and causing accidents. However, the power system also has immediate failures or the hidden danger trend is not obvious, so that the system fails to discover the hidden danger in time and suddenly appears a fault state. At this time, the present invention globally reconstructs the network topology map to eliminate the impact of the fault point and avoid the occurrence of a larger accident. The local reconstruction of the hidden danger state realizes the prevention of accidents in advance, and the rapid global reconstruction of the fault state realizes timely stop loss. The two-stage detection method and the strategy of taking different processing methods for different abnormalities guarantee the efficiency and flexibility of emergency operation and maintenance to a greater extent.
[0086] Furthermore, the step 1 determines the required number of batteries based on the load demand and operation and maintenance current of the power system to construct a parallel DC power supply, including:
[0087] Determining a first number of storage batteries required for the power system based on the rated total energy of the primary side of the power system and the rated capacity of a single storage battery;
[0088] Specifically, the load demand of the power system is measured by the rated total energy of the primary side of the power system, that is, the energy required for the normal operation of the entire power system.
[0089] The rated capacity of a single battery is the product of the rated voltage and rated current of the battery.
[0090] determining a second number of storage batteries required for the power system based on a maximum load current when an abnormality occurs in the power system and a rated output current of a single storage battery;
[0091] The operation and maintenance current of the power system is measured by the maximum load current when a fault occurs. It refers to the current required by the power system to at least maintain the normal operation of important equipment such as uninterruptible power supply equipment, important loads, protection and measurement and control devices when a fault occurs.
[0092] The larger value between the first number and the second number is the number of storage batteries required to construct a parallel DC power supply.
[0093] This embodiment determines the number of parallel power supply components through two aspects to ensure that the number of parallel batteries is sufficient for the entire system to continue to provide sufficient power in the early stage of an abnormal state, thereby fundamentally ensuring the emergency operation and maintenance capabilities of the power system.
[0094] In one embodiment, the determination of abnormal nodes in step 2 is achieved by establishing an abnormal location model, and each node in the network topology diagram is gradually searched according to the real-time operating parameter changes between nodes, and the nodes with sudden changes in real-time operating parameters are found to determine the abnormal nodes.
[0095] In one embodiment, see Figure 2 In step 3, the internal network of the power supply includes: in the parallel DC power supply, a plurality of single storage batteries, a switch array unit and a battery detection unit form an internal network of the power supply, the switch array unit controls the connection between the single storage batteries and the DC system bus to form a plurality of parallel battery modules, and the battery detection unit obtains the operating parameters of the parallel battery modules in real time.
[0096] When an abnormality occurs in the power system, the operating parameters of the battery detection unit are first obtained before searching for each node in the network topology diagram to determine whether the abnormal node is a power supply end. When it is determined that an abnormality occurs at the power supply end, the connection of the abnormal parallel battery module is disconnected, and under the constraint that the total output voltage and current of the parallel DC power supply change minimally, the connection mode and number of the remaining single batteries and the DC system bus are reconstructed, and the new switch array topology information is transmitted to the switch array unit to reconstruct the internal network of the power supply.
[0097] This embodiment utilizes the independent characteristics between parallel DC power supply units. Since the parallel single cells are isolated from each other, when an abnormality occurs in a single cell, it only affects the performance of the parallel battery module of this group without affecting other parallel battery modules. Therefore, there is no need to replace or process the entire power supply end, thereby reducing the cost and difficulty of power system operation and maintenance.
[0098] Preferably, the step 3 of determining the impact range of hidden dangers based on the real-time operating parameters of each node in the network topology diagram includes:
[0099] Obtaining the time point at which the hidden danger state of the power system occurs and the real-time operating parameters of the abnormal node at that time point, which are recorded as the hidden danger time point and the hidden danger parameter respectively;
[0100] Considering that the process of real-time operating parameters reaching the hidden danger threshold is often not sudden but continuous in the time dimension, that is, the hidden danger trend already exists for a period of time before reaching the hidden danger threshold, this embodiment determines the hidden danger impact range by analyzing the real-time operating parameters section by section to find the starting point of the hidden danger trend.
[0101] The real-time operating parameters of the abnormal node within a preset time period before the hidden danger time point are collected piece by piece at a preset time step and recorded as a parameter sequence;
[0102] The preset time period is obtained based on the duration of hidden danger trends determined during multiple operation and maintenance processes.
[0103] Based on the relative deviation between the parameter sequence and the hidden danger parameter, searching for the hidden danger trend starting point of the hidden danger state to determine the hidden danger trend duration;
[0104] In the network topology diagram, starting from the abnormal node according to the connection relationship between the nodes, the search for the first node within the impact range threshold is radially diffused outward;
[0105] The impact range threshold refers to the distance between the farthest node affected by the abnormal node and the abnormal node in the network topology diagram, and can be obtained based on the hidden danger impact range determined during multiple operation and maintenance processes.
[0106] The real-time operating parameters of the first node within the hidden danger trend duration and before the hidden danger trend starting point are respectively obtained, and the second node actually affected by the hidden danger state of the abnormal node is determined according to the difference value of the real-time operating parameters in the two time periods. The node set including all second nodes is the hidden danger impact range.
[0107] Preferably, searching for the starting point of the hidden danger trend of the hidden danger state based on the relative deviation between the parameter sequence and the hidden danger parameter to determine the hidden danger trend duration includes:
[0108] The hidden danger parameter and the characteristic values of the real-time operation parameter in each time period in the parameter sequence are extracted respectively; for example, the average value of each operation parameter in each time period can be calculated as the characteristic value of the real-time operation parameter.
[0109] Calculate the deviation value between the characteristic value of the real-time operating parameter and the characteristic value of the hidden danger parameter in each time period of the parameter sequence, and draw a deviation curve with time as the horizontal axis and the deviation value as the vertical axis;
[0110] The time point corresponding to the first inflection point of the deviation curve is the starting point of the hidden danger trend, and the time from the starting point of the hidden danger trend to the hidden danger time point is the duration of the hidden danger trend.
[0111] The deviation curve drawn in this embodiment reflects the trend that the operating parameters gradually approach the hidden danger parameters within the preset time period. As time changes, the deviation curve gradually decreases as a whole, that is, the difference between the operating parameters and the hidden danger parameters before the starting point of the hidden danger trend is large, and the difference between the operating parameters and the hidden danger parameters after the starting point of the hidden danger trend decreases rapidly. The inflection point of the curve refers to the connection point of the convex curve segment and the concave curve segment, which reflects the change of the operating trend or the operating rate. When the deviation curve has an inflection point, it indicates that the speed at which the operating parameters in the parameter sequence tend to the hidden danger parameters is accelerating, which is the starting point of the hidden danger trend.
[0112] For the first node, this embodiment is based on the experience value of the hidden danger impact range in multiple operation and maintenance processes. It is necessary to further screen the first node according to the real-time operating parameters before and after the starting point of the hidden danger trend to obtain the node actually affected in this hidden danger state. Specifically:
[0113] The real-time operating parameters of the first node within the hidden danger trend duration are plotted into a trend state curve with time as the horizontal axis and the real-time operating parameters as the vertical axis; the real-time operating parameters of the first node before the starting point of the hidden danger trend are plotted into a normal state curve with time as the horizontal axis and the real-time operating parameters as the vertical axis;
[0114] It should be understood that the trend state curve is the change in the real-time operating parameters of other nodes around the potential danger node after the potential danger trend appears, and it is compared with the real-time operating parameters of the corresponding node in the normal state. The more similar the trend state curve of a first node is to the normal state curve, that is, the more consistent the operating parameters of the first node are with the normal state within the potential danger trend of the potential danger node, indicating that the first node is still in a normal state, that is, the first node is not affected by the potential danger node.
[0115] A curve segment of the normal state curve having the same time length as the trend state curve and having the highest similarity is recorded as a comparison curve;
[0116] According to the above description, the method adopted in this embodiment is that if a curve consistent with the trend state curve can be found in the normal state curve, it means that the change of the first node during the period when the abnormal node generates a hidden danger trend is normal, that is, it is not affected by the hidden danger state.
[0117] The comparison curve is equally divided according to the preset time step in the trend state curve; the similarity between the trend state curve and the comparison curve in each time step is calculated segment by segment in chronological order, and the number of affected segments is counted: if the similarity is less than the similarity threshold, it means that the first node deviates from the normal state, that is, the segment may be affected by the potential danger node, and the number of affected segments is increased by one;
[0118] The first node whose number of affected segments is greater than a preset threshold is taken as a second node, and a node set consisting of all second nodes is recorded as the hidden danger impact range.
[0119] In this embodiment, the trend duration before the hidden danger state is determined by capturing the real-time operating parameter fluctuations of the abnormal node, and then the real-time operating parameters of other nodes that may be affected by the abnormal node before and after the hidden danger are compared to screen other nodes that are actually affected, and find the accurate distribution of nodes affected by the hidden danger to formulate corresponding emergency operation and maintenance strategies according to the actual degree of impact.
[0120] When an abnormal situation occurs in the power system, timely operation and maintenance are required to ensure normal operation. Under the premise of meeting various constraints of the system, the optimal opening and closing combination of all switches in the system can be found to reconstruct the network topology. Specifically, it includes:
[0121] Step 301, establishing a multi-objective optimization reconstruction model with one or more of minimum network loss, maximum load balancing, maximum power supply capacity, and minimum number of switch operations as objective functions, and one or more of voltage constraints, power constraints, and objective function weight constraints as constraint conditions;
[0122] In this embodiment, the objective function is , , , , They represent the minimum network loss objective function, the maximum load balancing objective function, the maximum power supply capacity objective function, and the minimum number of switch operations objective function, respectively. , , , They represent the weights of the corresponding objective functions respectively.
[0123] When the abnormal state is determined to be a hidden danger state and the network topology is partially reconstructed, the weight constraint of the objective function is: ;
[0124] When the abnormal state is determined to be a fault state and the network topology graph is globally reconstructed, the weight constraint of the objective function is: .
[0125] The expressions of the objective functions and other constraints are specifically set according to the actual scenarios and conditions of the power system, and are not limited in this embodiment.
[0126] Step 302, using a binary particle swarm optimization algorithm to find the optimal solution of the multi-objective optimization reconstruction model:
[0127] Different reconstruction schemes are abstracted into different particle positions; the dimension of the particle position is the number of switches in the network topology; the position of the particle in each dimension is 0 or 1, indicating the open or closed state of the corresponding switch in the network topology; the particle speed is defined as the probability of relative change of the particle position in the spatial dimension;
[0128] Constructing the initial particle swarm ,in is the number of reconstruction schemes, i.e., the number of particles; The particle at the 0th iteration location, , Represents particles In the The position in the dimensional space is also the The reconstruction scheme The state of the switch, Indicated in The reconstruction scheme A switch is disconnected, Indicated in The reconstruction scheme A switch is closed.
[0129] The relative change of a particle's position refers to the change of the particle's position relative to or tending towards a certain stable state, where "a certain stable state" can be the particle's position being 1 or the particle's position being 0 or a specific particle position at a certain moment.
[0130] Based on the objective function and the constraint conditions, calculating the fitness of each particle in the current particle swarm;
[0131] Iteratively update the particle speed and position to obtain the individual extreme value of each particle and the global extreme value of the particle group in each iteration process; the particle that obtains the individual extreme value / global extreme value is also the individual optimal particle / global optimal particle.
[0132] When the particle swarm converges or reaches the maximum number of iterations, the particle that currently obtains the global extreme value is output as the optimal solution of the multi-objective optimization reconstruction model, and the network topology graph is reconstructed according to the optimal solution.
[0133] In order to better meet the speed and cost requirements of emergency operation and maintenance, this embodiment adopts different processing strategies for abnormal situations of different degrees, including local load transfer to potential lines that have not yet caused substantial harm and global reconstruction of systems that have already failed.
[0134] When a fault has occurred in the power system, it is necessary to urgently cut off the faulty node to avoid a larger-scale fault. Because a certain node or a certain section of the line in the topology diagram is disconnected, the original power line is no longer a passage, and the entire system needs to reconstruct the network topology diagram for operation and maintenance. Specifically, a multi-objective optimization reconstruction model under a fault state is established, and the network topology diagram is globally reconstructed using the binary particle swarm optimization algorithm in the prior art. However, in the prior art, the binary particle swarm optimization algorithm has the problems of low computing power and slow convergence speed. This embodiment improves the speed update formula therein to meet the speed requirements of emergency operation and maintenance. Specifically including:
[0135] The abnormal nodes are removed from the network topology diagram, and an initial particle swarm is randomly generated. The speed of a particle is defined as the probability that the particle is at position 1 in each spatial dimension. The speed update formula is:
[0136]
[0137] in, Represents particles In the The fitness in the iteration; is the fitness dynamic factor, which is The fitness of the particle that obtains the global extreme value in the iteration The ratio of the average fitness of all particles in the iterations is negatively correlated; Represents particles In the In the iteration The speed of dimensional space; Represents particles In the In the iteration The speed of dimensional space; Represents particles When the individual extreme value is obtained, The position of the dimensional space; The particle that achieves the global extreme value is The position of the dimensional space; Represents particles In the In the iteration The position of the dimensional space; and is the learning factor; and A random number between 0 and 1.
[0138] In the binary particle swarm optimization algorithm, the probability of particles at a certain position is affected by changing the particle's speed to transform the particle position to the optimal one. In the speed update formula, The coefficient before represents the particle The speed inheritance is the weight of the speed of the particle in the previous iteration (for the sake of convenience, it is recorded as the speed inheritance coefficient in this embodiment). As the iteration proceeds, the particle position gradually tends to a stable optimal position. In the prior art, the speed inheritance coefficient is related to the number of iterations, but its change rule only follows the number of iterations and cannot adapt to the current actual situation of the particle. When the initial particle group is better, the convergence speed will be relatively slow, wasting a lot of time to ineffectively change the position of the particle. In this embodiment, the speed inheritance coefficient is adapted to the current fitness value of the particle to make improvements.
[0139] Specifically, due to the randomness of the initial population, there will be a large gap between the fitness of the global optimal particle at the beginning of the iteration and the average fitness of all current particles, that is, The value is small, the speed inheritance coefficient As a power function, The speed inheritance coefficient changes slowly as the value increases, that is, the speed change of this iteration is smaller than that of the previous iteration, and there will be more detailed speed change processes to search for a larger range of particle positions. In the later stages of the iteration, each particle is very close to the global optimal particle, so the fitness of the global optimal particle will be smaller than the average fitness of all current particles. The value of is large, the speed inheritance coefficient The change amplitude increases rapidly, so the speed in this iteration basically inherits the speed of the previous iteration and is smaller, and the probability of particle position change is very small, which makes the algorithm converge. Rapid convergence to the optimal particle and faster acquisition of the optimal reconstruction solution realizes the rapid and efficient emergency operation and maintenance.
[0140] On the other hand, in this embodiment, the operating parameters of each node are measured in real time, which can timely discover the hidden danger states before most faults. When the abnormal state of the power system is a hidden danger state, no fault has occurred at this time, but there is a fault trend that needs to be dealt with in time. This embodiment considers transferring the line load with hidden dangers to other nodes and lines, that is, focusing on optimizing the topological structure within the hidden danger range, and the structure outside the hidden danger range is slightly changed to adapt to the reconstruction processing within the hidden danger range. Specifically, when the abnormal state is determined to be a hidden danger state, a multi-objective optimization reconstruction model under the hidden danger state is established, and the binary particle swarm optimization algorithm is used to locally reconstruct the network topology map based on the hidden danger impact range, including:
[0141] The topological structure in the network topological diagram outside the impact range of the hidden danger at the time point of the hidden danger is fixed, and the topological structure in the network topological diagram within the impact range of the hidden danger is randomly generated to form an initial particle swarm. That is, reconstruction is performed based on the network topological structure when the hidden danger occurs. Specifically, the state of the switch between the nodes outside the impact range of the hidden danger when the hidden danger occurs (where the position value is still 0 or 1, and the particle corresponding to the closed switch in the network topological structure when the hidden danger occurs is 1, and the switch is disconnected is 0) is recorded. This part does not change much during the reconstruction process; the node position affected by the hidden danger node is randomly generated, and this part needs to be replanned to find a better structure. The two parts are combined to generate the initial particle swarm.
[0142] The existing particle swarm algorithm has a reconstruction scheme that initializes the particle swarm with a large number of particle positions being 0, that is, the switch is disconnected, and then changes the particle speed to gradually change the particle position to 1, that is, the switch is closed. However, this application considers local power transfer in a hidden danger state, in which a large number of topological structures that are already optimal do not need to be changed significantly. At this time, it is unreasonable to use the conventional technology to define the particle speed as the probability of the particle position being 1 and continuously increase this scheme, so the particle speed is defined as the probability of the particle position changing in each spatial dimension;
[0143] For example Represents particles In the The probability of a particle position changing in the dimensional space, that is, The reconstruction scheme The probability that the state of a switch changes relative to the initial state of the switch in the scheme.
[0144] When the local reconstruction is performed in this embodiment, the particle speed is determined by the impact range of the hidden danger.
[0145] Specifically, within the range of hidden danger influence, as the center of the hidden danger node is approached, the switch (particle position) needs to change relative to the original state. Therefore, the probability of change of the switch state is weighted based on the distance between each switch and the abnormal node, that is, the speed of the particle in each spatial dimension is weighted. The speed update formula is:
[0146] If the switch is within the range of the hidden danger,
[0147]
[0148] is the distance factor, , is negatively correlated with the distance between the switch and the abnormal node, indicating that the farther the distance between the switch and the abnormal node, the smaller the probability of the switch state changing.
[0149] Outside the range of hidden dangers, the topological structure is close to optimal, and the probability of switch state change is small, which is only determined by the current global optimal particle position.
[0150] If the switch is outside the scope of the hidden danger,
[0151]
[0152] in, Represents particles In the In the iteration The speed of the space, Represents particles In the In the iteration The speed of the space, Represents particles When the individual extreme value is obtained, The position of the dimensional space, The particle that achieves the global extreme value is The position of the dimensional space, Represents particles In the In the iteration The position of the dimensional space, and is the learning factor, and is a random number between 0 and 1;
[0153] It is the inertia factor, which indicates the proportion of the particle's velocity in this iteration inherited from the velocity of the same particle in the previous iteration. It is negatively correlated with the number of iterations, indicating that as the number of iterations increases, the probability of particle position change decreases.
[0154] In the above reconstruction process, the positions of particles within the hidden danger influence range in the hidden danger state and in the fault state are updated using the following formula:
[0155]
[0156] in, It is a random number between 0 and 1. The meanings of the other parameters are the same as those in the particle speed update and will not be repeated here. Used to map the particle's velocity to the range [0,1].
[0157] In the hidden danger state, the position of particles outside the hidden danger range is updated using the following formula:
[0158]
[0159] in, , Represents particles In the , The fitness value in the iteration, Represents particles In the The meanings of the other parameters are the same as above and will not be repeated here.
[0160] In this embodiment, according to the different impacts of different abnormal situations on the network topology of the power system, emergency operation and maintenance plans with different scopes and different reconstruction targets are adopted, and load transfer to loads that may fail is performed in advance based on the scope of the affected nodes to reduce the possibility of accidents. At the same time, small-scale reconstruction also reduces operation and maintenance costs.
[0161] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0162] Based on the same inventive concept, the embodiment of the present application also provides an operating system for implementing the above-mentioned power system emergency operation and maintenance method based on parallel DC power supply, such as Figure 3 As shown, including:
[0163] A parallel DC power supply building module is used to build a parallel DC power supply based on the load demand of the power system and the number of batteries required for operation and maintenance current calculation;
[0164] A topology diagram building module is used to build a network topology diagram corresponding to the power system according to the connection relationship between the power supply end and each load end in the power system;
[0165] An abnormal node determination module is used to collect real-time operating parameters of each node in the network topology diagram to determine abnormal nodes;
[0166] The emergency operation and maintenance module is used to select a network reconstruction scheme according to the location of the abnormal node to perform emergency operation and maintenance of the power system, including:
[0167] When the abnormal node is a power supply end, changing the connection mode of the single battery in the parallel DC power supply to reconstruct the internal network of the power supply;
[0168] When the abnormal node is a load end, determining the abnormal state type of the power system based on the real-time operating parameters, the abnormal state type includes a hidden danger state and a fault state;
[0169] When the abnormal state is determined to be a hidden danger state, determining the hidden danger impact range based on the real-time operating parameters of each node in the network topology map, and partially reconstructing the network topology map based on the hidden danger impact range;
[0170] When the abnormal state is determined to be a fault state, the abnormal node is removed and the network topology diagram is globally reconstructed to implement emergency operation and maintenance.
[0171] The implementation solution provided by the device to solve the problem is the same as the implementation solution recorded in the above method. Therefore, the specific limitations of the implementation example of the implementation system provided below can be found in the above limitations on the emergency operation and maintenance method of the power system based on a parallel DC power supply, and will not be repeated here.
[0172] The present invention is not limited to the above-mentioned specific implementation modes. Various changes made by ordinary technicians in this field based on the above-mentioned concepts without creative work are all within the protection scope of the present invention.
Claims
1. An emergency operation and maintenance method for a power system based on a parallel DC power supply, characterized in that: The steps include: Step 1: Determine the number of batteries required to build a parallel DC power supply based on the load demand and operation and maintenance current of the power system; Step 2, constructing a network topology diagram corresponding to the power system according to the connection relationship between the power supply end and each load end in the power system, collecting real-time operating parameters of each node in the network topology diagram and determining abnormal nodes; Step 3, when the abnormal node is a power supply end, changing the connection mode of the single battery in the parallel DC power supply to reconstruct the internal network of the power supply; When the abnormal node is a load end, determining the abnormal state type of the power system based on the real-time operating parameters, the abnormal state type includes a hidden danger state and a fault state; When the abnormal state is determined to be a hidden danger state, determining the hidden danger impact range based on the real-time operating parameters of each node in the network topology map, and partially reconstructing the network topology map based on the hidden danger impact range; When the abnormal state is determined to be a fault state, the abnormal node is removed and the network topology is globally reconstructed to implement emergency operation and maintenance; Wherein, the step 3 of determining the impact scope of hidden dangers based on the real-time operating parameters of each node in the network topology diagram includes: Obtaining the time point at which the hidden danger state of the power system occurs and the real-time operating parameters of the abnormal node at that time point, which are recorded as the hidden danger time point and the hidden danger parameter respectively; The real-time operating parameters of the abnormal node within a preset time period before the hidden danger time point are collected piece by piece at a preset time step and recorded as a parameter sequence; Based on the relative deviation between the parameter sequence and the hidden danger parameter, searching for the hidden danger trend starting point of the hidden danger state to determine the hidden danger trend duration; In the network topology diagram, starting from the abnormal node according to the connection relationship between the nodes, the search for the first node within the impact range threshold is radially diffused outward; The real-time operating parameters of the first node within the hidden danger trend duration and before the hidden danger trend starting point are respectively obtained, and the second node actually affected by the hidden danger state of the abnormal node is determined according to the difference value of the real-time operating parameters in the two time periods, and the node set including all the second nodes is the hidden danger impact range; The step of searching for a hidden danger trend starting point of a hidden danger state based on a relative deviation between the parameter sequence and the hidden danger parameter to determine the hidden danger trend duration includes: Respectively extracting the hidden danger parameter and the characteristic values of the real-time operation parameter in each time period in the parameter sequence; Calculate the deviation value between the characteristic value of the real-time operating parameter and the characteristic value of the hidden danger parameter in each time period of the parameter sequence, and draw a deviation curve with time as the horizontal axis and the deviation value as the vertical axis; The time point corresponding to the first inflection point of the deviation curve is the starting point of the hidden danger trend, and the time from the starting point of the hidden danger trend to the hidden danger time point is the duration of the hidden danger trend; The determining of the second node actually affected by the hidden danger state of the abnormal node according to the difference value of the real-time operating parameters in two time periods, and the node set including all the second nodes being the hidden danger impact range, includes: For each of the first nodes, The real-time operating parameters of the first node within the hidden danger trend duration are plotted into a trend state curve with time as the horizontal axis and the real-time operating parameters as the vertical axis; the real-time operating parameters of the first node before the starting point of the hidden danger trend are plotted into a normal state curve with time as the horizontal axis and the real-time operating parameters as the vertical axis; A curve segment of the normal state curve having the same time length as the trend state curve and having the highest similarity is recorded as a comparison curve; The comparison curve is equally divided according to the preset time step in the trend state curve; the similarity between the trend state curve and the comparison curve in each time step is calculated segment by segment in chronological order, and the number of affected segments is counted: if the similarity is less than a similarity threshold, the number of affected segments is increased by one; The first node whose number of affected segments is greater than a preset threshold is taken as a second node, and a node set consisting of all second nodes is recorded as the hidden danger impact range.
2. The method for emergency operation and maintenance of a power system based on a parallel DC power supply according to claim 1, characterized in that: Step 1 determines the number of batteries required to build a parallel DC power supply based on the load demand and operation and maintenance current of the power system, including: Determining a first number of storage batteries required for the power system based on the rated total energy of the primary side of the power system and the rated capacity of a single storage battery; determining a second number of storage batteries required for the power system based on a maximum load current when an abnormality occurs in the power system and a rated output current of a single storage battery; The larger value between the first number and the second number is the number of storage batteries required to construct a parallel DC power supply.
3. The method for emergency operation and maintenance of a power system based on a parallel DC power supply according to claim 1, characterized in that: Step 3, changing the connection mode of the single battery in the parallel DC power supply to reconstruct the internal network of the power supply, includes: Disconnect the abnormal battery, and under the constraint that the total output voltage and current of the parallel DC power supply change minimally, change the connection mode and number of the remaining single batteries and the DC system bus to reconstruct the internal network of the power supply.
4. The method for emergency operation and maintenance of a power system based on a parallel DC power supply according to claim 1, characterized in that: The reconstructing the network topology diagram specifically includes: A multi-objective optimization reconstruction model is established with one or more of minimum network loss, maximum load balancing, maximum power supply capacity, and minimum number of switch operations as objective functions, and one or more of voltage constraints, power constraints, and objective function weight constraints as constraint conditions; The binary particle swarm optimization algorithm is used to solve the optimal solution of the multi-objective optimization reconstruction model: Different reconstruction schemes are abstracted into different particle positions; the dimension of the particle position is the number of switches in the network topology; the position of the particle in each dimension is 0 or 1, indicating the open or closed state of the corresponding switch in the network topology; the particle speed is defined as the probability of relative change of the particle position in the spatial dimension; Based on the objective function and the constraint conditions, calculating the fitness of each particle in the current particle swarm; Iteratively update the speed and position of particles to obtain the individual extreme value of each particle and the global extreme value of the particle group in each iteration process; When the particle swarm converges or reaches the maximum number of iterations, the particle that currently obtains the global extreme value is output as the optimal solution of the multi-objective optimization reconstruction model, and the network topology graph is reconstructed according to the optimal solution.
5. The method for emergency operation and maintenance of a power system based on a parallel DC power supply according to claim 4, characterized in that: When the abnormal state is determined to be a hidden danger state, a multi-objective optimization reconstruction model under the hidden danger state is established, and a binary particle swarm optimization algorithm is used to locally reconstruct the network topology map based on the hidden danger impact range, specifically including: The topological structure in the network topological map outside the hidden danger impact range at the hidden danger time point is fixed, and the topological structure in the network topological map within the hidden danger impact range is randomly generated to form an initial particle group; the speed of the particle is defined as the probability of the particle position changing in each spatial dimension; The probability of the switch state change is weighted based on the distance between each switch and the abnormal node, that is, the speed of the particle in each spatial dimension is weighted. The speed update formula is: If the switch is within the range of the hidden danger, ; If the switch is outside the range of the hidden danger, ; in, Represents particles In the In the iteration The speed of the space, Represents particles In the In the iteration The speed of the space, Represents particles When the individual extreme value is obtained, The position of the dimensional space, The particle that achieves the global extreme value is The position of the dimensional space, Represents particles In the In the iteration The position of the dimensional space, and is the learning factor, and is a random number between 0 and 1; is the inertia factor, which is negatively correlated with the number of iterations, indicating that as the number of iterations increases, the probability of particle position change decreases; is the distance factor, , is negatively correlated with the distance between the switch and the abnormal node, indicating that the farther the distance between the switch and the abnormal node, the smaller the probability of the switch state changing.
6. The method for emergency operation and maintenance of a power system based on a parallel DC power supply according to claim 4, characterized in that: When the abnormal state is determined to be a fault state, a multi-objective optimization reconstruction model under the fault state is established, and a binary particle swarm optimization algorithm is used to globally reconstruct the network topology graph, specifically including: The abnormal nodes are removed from the network topology diagram, and an initial particle swarm is randomly generated. The speed of a particle is defined as the probability that the particle is at position 1 in each spatial dimension. The speed update formula is: ; in, Represents particles In the The fitness in the iteration; is the fitness dynamic factor, which is The fitness of the particle that obtains the global extreme value in the iteration The ratio of the average fitness of all particles in the iterations is negatively correlated; Represents particles In the In the iteration The speed of dimensional space; Represents particles In the In the iteration The speed of dimensional space; Represents particles When the individual extreme value is obtained, The position of the dimensional space; The particle that achieves the global extreme value is The position of the dimensional space; Represents particles In the In the iteration The position of the dimensional space; and is the learning factor; and A random number between 0 and 1.
7. An emergency operation and maintenance system for a power system based on a parallel DC power supply, characterized in that: include: A parallel DC power supply building module is used to build a parallel DC power supply based on the load demand of the power system and the number of batteries required for operation and maintenance current calculation; A topology diagram building module is used to build a network topology diagram corresponding to the power system according to the connection relationship between the power supply end and each load end in the power system; An abnormal node determination module is used to collect real-time operating parameters of each node in the network topology diagram to determine abnormal nodes; The emergency operation and maintenance module is used to select a network reconstruction scheme according to the location of the abnormal node to perform emergency operation and maintenance of the power system, including: When the abnormal node is a power supply end, changing the connection mode of the single battery in the parallel DC power supply to reconstruct the internal network of the power supply; When the abnormal node is a load end, determining the abnormal state type of the power system based on the real-time operating parameters, the abnormal state type includes a hidden danger state and a fault state; When the abnormal state is determined to be a hidden danger state, determining the hidden danger impact range based on the real-time operating parameters of each node in the network topology map, and partially reconstructing the network topology map based on the hidden danger impact range; When the abnormal state is determined to be a fault state, the abnormal node is removed and the network topology is globally reconstructed to implement emergency operation and maintenance; The emergency operation and maintenance module determines the impact range of hidden dangers based on the real-time operation parameters of each node in the network topology diagram, including: Obtaining the time point at which the hidden danger state of the power system occurs and the real-time operating parameters of the abnormal node at that time point, which are recorded as the hidden danger time point and the hidden danger parameter respectively; The real-time operating parameters of the abnormal node within a preset time period before the hidden danger time point are collected piece by piece at a preset time step and recorded as a parameter sequence; Based on the relative deviation between the parameter sequence and the hidden danger parameter, searching for the hidden danger trend starting point of the hidden danger state to determine the hidden danger trend duration; In the network topology diagram, starting from the abnormal node according to the connection relationship between the nodes, the search for the first node within the impact range threshold is radially diffused outward; The real-time operating parameters of the first node within the hidden danger trend duration and before the hidden danger trend starting point are respectively obtained, and the second node actually affected by the hidden danger state of the abnormal node is determined according to the difference value of the real-time operating parameters in the two time periods, and the node set including all the second nodes is the hidden danger impact range; The step of searching for a hidden danger trend starting point of a hidden danger state based on a relative deviation between the parameter sequence and the hidden danger parameter to determine the hidden danger trend duration includes: Respectively extracting the hidden danger parameter and the characteristic values of the real-time operation parameter in each time period in the parameter sequence; Calculate the deviation value between the characteristic value of the real-time operating parameter and the characteristic value of the hidden danger parameter in each time period of the parameter sequence, and draw a deviation curve with time as the horizontal axis and the deviation value as the vertical axis; The time point corresponding to the first inflection point of the deviation curve is the starting point of the hidden danger trend, and the time from the starting point of the hidden danger trend to the hidden danger time point is the duration of the hidden danger trend; The determining of the second node actually affected by the hidden danger state of the abnormal node according to the difference value of the real-time operating parameters in two time periods, and the node set including all the second nodes being the hidden danger impact range, includes: For each of the first nodes, The real-time operating parameters of the first node within the hidden danger trend duration are plotted into a trend state curve with time as the horizontal axis and the real-time operating parameters as the vertical axis; the real-time operating parameters of the first node before the starting point of the hidden danger trend are plotted into a normal state curve with time as the horizontal axis and the real-time operating parameters as the vertical axis; A curve segment of the normal state curve having the same time length as the trend state curve and having the highest similarity is recorded as a comparison curve; The comparison curve is equally divided according to the preset time step in the trend state curve; the similarity between the trend state curve and the comparison curve in each time step is calculated segment by segment in chronological order, and the number of affected segments is counted: if the similarity is less than a similarity threshold, the number of affected segments is increased by one; The first node whose number of affected segments is greater than a preset threshold is taken as a second node, and a node set consisting of all second nodes is recorded as the hidden danger impact range.
