A method and system for power system simulation based on Bayesian network

Abstract

The present invention discloses a method and system for simulating an electric power system based on a Bayesian network, and relates to the technical field of electric power systems. The method for simulating an electric power system based on a Bayesian network comprises the following steps: constructing a Bayesian network model; confidence assessment; model training; and establishing a simulation flow circuit. The present invention collects key data of the electric power system within a preset time period and constructs a Bayesian network model, then monitors the operating status of the electric power system in real time and performs confidence assessment, then trains the Bayesian network model according to the result of the confidence assessment, and finally establishes a simulation flow circuit of the electric power system based on the Bayesian network model after model training, and simultaneously monitors the changing trend of the voltage value and the current value on the power line of the simulation flow circuit in real time during the simulation process, thereby achieving the effect of improving the simulation accuracy of the electric power system, and solving the problem of low correlation between the simulation accuracy of the electric power system and the real-time operating status of the electric power system in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a power system simulation method and system based on a Bayesian network. Background Art

[0002] With the continuous development and expansion of the power system, the number of devices and components contained in the system has increased dramatically, resulting in an increase in the complexity and uncertainty of the system. Secondly, with the continuous development of big data and artificial intelligence technology, the power system simulation method based on Bayesian network will be more widely used and deeply studied. As a probabilistic graphical model, Bayesian network can accurately model and infer complex relationships, and is particularly suitable for dealing with uncertainty problems. It combines the advantages of probability theory and graph theory, describes uncertainty problems by expressing conditional dependencies between variables, and provides new ideas and methods for power system simulation. In addition, power system simulation is of great significance in ensuring the reliability of power supply, optimizing system operation, and improving energy utilization efficiency. However, traditional power system simulation methods have limitations in dealing with complexity and uncertainty, and it is difficult to accurately describe and predict the behavior of the system. Therefore, in order to improve the intelligence and sustainable development of the power system, a power system simulation method and system based on Bayesian network came into being.

[0003] The existing power system simulation method generates a corresponding power grid data set by randomly sampling the power system in the target area, and then simulates and analyzes the power grid data set according to the power grid sample set and constructs a simulation analysis model. The constructed simulation analysis model is then used to predict the stability of the power system during real-time operation.

[0004] For example, the invention patent with announcement number: CN113839385B announces a method for estimating inertia of a power system based on Bayesian estimation, which includes: obtaining model parameters, and making prior assumptions on the model parameters based on Bayesian estimation theory to obtain a priori probability distribution; using observed data to obtain a likelihood function, and using the Bayesian formula to calculate the likelihood function and the prior probability distribution to obtain a posterior distribution; based on the posterior distribution, online estimation is used to complete inertia evaluation.

[0005] For example, a patent application with publication number: CN116780548A discloses a credibility assessment method and device for the voltage support strength of an electric power system, comprising: obtaining real-time measurement data of the electric power system to be tested and inputting it into a pre-trained critical short-circuit ratio prediction model, and outputting a first critical short-circuit ratio interval; optimizing the first critical short-circuit ratio interval using a Bayesian algorithm to determine a second critical short-circuit ratio interval; determining a critical short-circuit ratio prediction value of the electric power system to be tested from the second critical short-circuit ratio interval, obtaining a real-time new energy multi-station short-circuit ratio of the electric power system to be tested, and subtracting it from the critical short-circuit ratio prediction value to determine the real-time voltage support strength margin of the electric power system to be tested.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0007] In the prior art, since the power grid data set in the target area is continuously updated during the generation process, the complexity of the sample data set during the power system simulation increases, and there is a problem of low correlation between the power system simulation accuracy and the real-time operating status of the power system. Summary of the invention

[0008] The embodiment of the present application solves the problem of low correlation between power system simulation accuracy and real-time operating status of the power system in the prior art by providing a power system simulation method and system based on a Bayesian network, thereby improving the accuracy of power system simulation.

[0009] An embodiment of the present application provides a method for simulating an electric power system based on a Bayesian network, comprising the following steps: S1, collecting key data of the electric power system within a preset time period and constructing a Bayesian network model; S2, monitoring the operating status of the electric power system in real time and performing confidence assessment based on the conditional probability distribution of the key data in the Bayesian network model; S3, obtaining the grid data of the electric power system during operation within a preset time period and training the Bayesian network model based on the result of the confidence assessment; S4, establishing a simulated power flow circuit of the electric power system based on the Bayesian network model after model training, and simultaneously monitoring in real time the changing trends of the voltage and current values ​​on the power lines in the simulated power flow circuit during the simulation process.

[0010] Furthermore, the specific construction steps of the Bayesian network model are: standardizing the collected key data and mapping the standardized key data to the corresponding node positions in the Bayesian network in the form of node variables according to the preset node variables in the power system; obtaining the node matching index and constructing a directed acyclic graph according to the actual operating principle of the power system and the dependency relationship between the power system structure and the nodes, and the node matching index is used to quantify the spatial matching degree between the preset node variables in the power system and the node positions in the Bayesian network; setting the Bayesian network structure according to the constructed directed acyclic graph and constructing the Bayesian network model in combination with the key nodes in the power system.

[0011] Furthermore, the node matching index is obtained by the following method: inputting the key data after data standardization into the electrical wiring diagram of the power system to establish an electrical connection relationship diagram between the node and the boundary, and monitoring the change of the electrical connection distance of the electrical components in a preset time period in real time to obtain a distance measurement index; obtaining an angle closeness coefficient according to the angle between the node and the boundary in the electrical connection relationship diagram and the actual needs of the power system, and obtaining the functional similarity according to the functional characteristics and electrical connection angle threshold between the nodes in the electrical connection relationship diagram combined with the electrical function similarity score corresponding to the node in the preset time period, wherein the electrical connection angle threshold represents the angle range between the electrical components obtained after weighted summation based on the historical operation data of the power system; collecting the electrical function data of the nodes in the preset time period, adjusting the degree of fluctuation deviation in the electrical function data according to the actual simulation needs of the power system, and obtaining a preset function similarity correction factor from the database, and combining the obtained angle closeness coefficient and function similarity to obtain the node matching index, wherein the electrical function data includes voltage fluctuation, current fluctuation and power output, and the function similarity correction factor is used to correct the degree of fluctuation deviation of the electrical function data during the operation of the power system.

[0012] Furthermore, the node matching index is calculated by the following formula:

[0013]

[0014] Where r is the number of the electrical component, r = 1, 2, ..., R, R is the total number of electrical components, m is the number of the preset time period, m = 1, 2, ..., M, M is the total number of preset time periods, JIE r.m It represents the node matching index of the rth electrical component in the mth preset time period, e is a natural constant, A r.m represents the distance measurement index of the rth electrical component in the mth preset time period, A0 represents the distance reference measurement index, ΔA represents the distance measurement index reference deviation, B r.mrepresents the functional similarity of the rth electrical component in the mth preset time period, α represents the functional similarity correction factor, and θ r.m represents the angle tightness coefficient of the rth electrical component in the mth preset time period, and δ represents the electrical connection angle threshold.

[0015] Furthermore, the specific steps of establishing a simulated power flow circuit of the power system based on the Bayesian network model after model training are: setting the initial parameters of the simulated power flow circuit according to the actual simulation requirements of the power system, and dynamically adjusting the initial parameters according to the results of model training, the initial parameters including generator output and power line impedance; real-time monitoring of the conditional dependency between electrical components during the dynamic adjustment of the initial parameters, and obtaining a correlation influence index by visualizing the influence path between electrical components according to the Bayesian network model, the correlation influence index being used to measure the degree of influence of electrical components on the simulated power flow circuit within a preset time period; simulating the operating state of the Bayesian network model in a preset scenario according to the influence path and correlation influence index between electrical components and in combination with the actual circuit of the power system to construct a simulated power flow circuit, and using the output of the Bayesian network model as the input of the simulated power flow circuit to reflect the actual operating state of the power system and the power flow distribution of the simulated power flow circuit in real time.

[0016] Furthermore, the specific steps for obtaining the correlation influence index are: collecting the starting value and the ending value of the initial parameters of the power system during operation under the preset scenario, and monitoring the numerical changes of the electrical components in the process of dynamically adjusting the initial parameters in real time to obtain the electrical influence coefficient and obtain the preset initial parameter correction factor from the preset database; real-time monitoring the influence of the electrical components of the simulated power flow circuit on the power flow distribution path under the preset scenario to obtain the distribution deviation rate and dynamically adjust the power flow distribution path according to the preset path of the power flow distribution, and obtain the preset dynamic adjustment factor from the preset database; combining the obtained electrical influence coefficient and the initial parameter correction factor to obtain the correlation influence index, and the correlation influence index is calculated by the following formula:

[0017]

[0018] In the formula, r is the number of the electrical component, r = 1, 2, ..., R, R is the total number of electrical components, n is the number of the preset scene, n = 1, 2, ..., N, N is the total number of preset scenes, GU r.n represents the correlation impact index of the rth electrical component in the nth preset scenario, e is a natural constant, P r 2 .n represents the initial parameter ending value of the rth electrical component in the nth preset scenario, Pr 1 .n represents the initial parameter starting value of the rth electrical component in the nth preset scenario, ΔP represents the initial parameter reference deviation, D r.n It represents the electrical influence coefficient of the rth electrical component in the nth preset scenario, Q r.n It represents the distribution deviation rate of the rth electrical component in the nth preset scenario, Q0 represents the reference distribution deviation rate, γ represents the dynamic adjustment factor, and β represents the initial parameter correction factor.

[0019] An embodiment of the present application provides an artificial intelligence-based SMS content analysis system, including: a Bayesian network model construction module, a confidence assessment module, a model training module and a simulation flow circuit establishment module; wherein the Bayesian network model construction module is used to collect key data of the power system within a preset time period and construct a Bayesian network model; the confidence assessment module is used to monitor the operating status of the power system in real time and perform confidence assessment based on the conditional probability distribution of the key data in the Bayesian network model; the model training module is used to obtain the power grid data of the power system during operation within a preset time period and perform model training on the Bayesian network model according to the result of the confidence assessment; the simulation flow circuit establishment module is used to establish a simulation flow circuit of the power system based on the Bayesian network model after model training, and at the same time monitor in real time the changing trend of the voltage and current values ​​on the power line of the simulation flow circuit during the simulation process.

[0020] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0021] 1. By collecting key data of the power system within a preset time period and constructing a Bayesian network model, the operating status of the power system is monitored in real time and a confidence assessment is performed. Then, the Bayesian network model is trained according to the results of the confidence assessment. Finally, a simulated power flow circuit of the power system is established based on the Bayesian network model after model training. At the same time, the changing trend of the voltage and current values ​​on the power line during the simulation process is monitored in real time in the simulated power flow circuit, thereby achieving a more accurate establishment of the simulated power flow circuit, and then achieving an improvement in the accuracy of power system simulation, effectively solving the problem of low correlation between the power system simulation accuracy and the real-time operating status of the power system in the prior art.

[0022] 2. The initial condition probability distribution of the nodes in the Bayesian network model is set according to the actual simulation requirements of the power system. At the same time, the load data of the power system in a preset time period is collected in real time and input into the Bayesian network model as training data. Then, the parameters of the Bayesian network model are initialized and the Bayesian network model is trained using the training data. At the same time, the model training loss function is defined according to the performance of the power system during operation, and the Bayesian network model is iteratively optimized in combination with the gradient descent algorithm, thereby achieving more accurate optimization of the Bayesian network model and improving the accuracy and reliability of the Bayesian network model.

[0023] 3. Set the initial parameters of the simulated power flow circuit according to the actual simulation requirements of the power system, dynamically adjust the initial parameters according to the results of model training, and monitor the conditional dependencies between electrical components in the process of dynamic adjustment of the initial parameters in real time. Then, combine the actual circuit of the power system to simulate the operating status of the Bayesian network model in the preset scenario to construct a simulated power flow circuit. Finally, use the output of the Bayesian network model as the input of the simulated power flow circuit to reflect the actual operating status of the power system and the power flow distribution of the simulated power flow circuit in real time, thereby improving the accuracy of the construction of the simulated power flow circuit and further improving the reliability of the power system simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flow chart of a power system simulation method based on a Bayesian network provided in an embodiment of the present application;

[0025] Figure 2 A flow chart for establishing a simulation power flow circuit provided in an embodiment of the present application;

[0026] Figure 3 A three-dimensional statistical analysis chart of the correlation impact index provided in the embodiment of the present application;

[0027] Figure 4 A schematic diagram of the structure of a power system simulation system based on a Bayesian network provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The embodiment of the present application solves the problem of low accuracy of power system simulation in the prior art by providing a power system simulation method and system based on a Bayesian network. The key data of the power system within a preset time period is collected and a Bayesian network model is constructed through a Bayesian network model construction module. Then, the operating status of the power system is monitored in real time and a confidence assessment is performed according to the conditional probability distribution of the key data in the Bayesian network model through a confidence assessment module. Then, the power grid data of the power system during operation within a preset time period is obtained through a model training module. At the same time, the output of the generator and the load distribution of the power system are input into the Bayesian network model and trained according to the result of the confidence assessment. Finally, a simulation power flow circuit of the power system is established based on the Bayesian network model after model training through a simulation power flow circuit establishment module. At the same time, the changing trends of the voltage and current values ​​on the power line during the simulation of the simulation power flow circuit are monitored in real time, thereby improving the accuracy of power system simulation.

[0029] The technical solution in the embodiment of the present application is to solve the problem that the accuracy of the power system simulation is low in correlation with the real-time operating state of the power system. The overall idea is as follows:

[0030] By collecting key data of the power system within a preset time period and constructing a Bayesian network model, the operating status of the power system is monitored in real time and a confidence assessment is performed. The Bayesian network model is then trained according to the results of the confidence assessment. Finally, a simulated power flow circuit of the power system is established based on the trained Bayesian network model. At the same time, the changing trends of the voltage and current values ​​on the power lines in the simulated power flow circuit are monitored in real time during the simulation process, thereby achieving the effect of improving the simulation accuracy of the power system.

[0031] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0032] like Figure 1As shown, it is a flow chart of a power system simulation method based on a Bayesian network provided in an embodiment of the present application, the method comprising the following steps: S1, collecting key data of the power system within a preset time period and constructing a Bayesian network model, the key data including power supply voltage data, load change data and transmission circuit data, the Bayesian network model is used to predict the load change of the power system within a preset time period; S2, monitoring the operating state of the power system in real time and performing confidence evaluation according to the conditional probability distribution of the key data in the Bayesian network model, the conditional probability distribution is used to describe the probability of a circuit failure in the power system under preset voltage conditions, and the confidence evaluation is used to quantify the credibility of the conditional probability distribution result; S3, obtaining the grid data of the power system during operation within a preset time period and training the Bayesian network model according to the confidence assessment results. The grid data represents the real-time operating status of the power system under a preset load. Model training represents inputting the generator output and the load distribution of the power system into the Bayesian network model and training the model according to the confidence assessment results. S4, establishing a simulated power flow circuit of the power system based on the Bayesian network model after model training, and monitoring the changing trends of the voltage and current values ​​on the power lines in the simulated power flow circuit in real time during the simulation process. The simulated power flow circuit is used to visualize the influence of electrical components in the power system on the power flow distribution.

[0033] In this embodiment, the supply voltage data generally includes voltage values ​​and voltage fluctuations at different time points; the load change data generally indicates load changes within a preset time period, such as load increase and decrease trends, peak loads, and average loads; the transmission circuit data generally indicates relevant data of the transmission circuit, such as transmission efficiency, line loss, and current changes; the conditional probability distribution generally describes the probability distribution under a given node state; for each node, there is a conditional probability table associated with it; for example, when evaluating the state of a generator node in a power system, the probability of the generator being in a normal or faulty state can be determined by referring to the conditional probability table corresponding to the node based on the states of other components connected to it (such as a transformer); voltage nodes are generally used to describe the voltage level in a power system, and take into account the voltage fluctuation over time; current nodes generally indicate the current intensity in a power system, and take into account its changes over time; Rate nodes are usually used to describe the conversion rate of energy in the power system, and take into account the difference between active power and reactive power; for each node variable, its numerical range and quantization level need to be determined. For example, the voltage may fluctuate within a specific range (such as 220V±10%), which can be divided into several levels (such as low, medium, and high). If the voltage value is between (198V) and (210V), it is judged as a low-level voltage; if the voltage value is between (210V) and (230V), it is judged as a medium-level voltage; if the voltage value is between (230V) and (242V), it is judged as a high-level voltage; then specific values ​​are assigned to each node variable based on the actual operating data of the power system. These values ​​are usually obtained by weighted averaging based on historical data. For example, if the voltage at a certain point in time is 230V and is at a medium level, then the value of the voltage node can be set to the corresponding value of the medium level.

[0034] Specifically, power grid data is usually used to describe real-time operating status information related to generator output, load distribution, voltage, current and power, so as to reflect the actual operating conditions of the power system. The generator output and the load distribution of the power system are adjusted according to the results of the confidence assessment. The adjustment can be based on the actual needs, operating strategies and optimization goals of the power system. For example, if the load in a certain area is too high, it is necessary to increase the generator output in that area or redistribute the load. The parameters of each component in the simulation flow circuit are set according to the actual parameters of the power system. During the power system simulation process, the real-time data is passed to the Bayesian network model. By monitoring the changing trends of the voltage and current values ​​on the power lines in real time, the changing laws of the voltage and current on the power lines are understood, and the changing laws are displayed in the form of graphs. It should be noted that with the changes in the power system (equipment replacement and network structure changes), the simulation flow circuit needs to be updated regularly, and the time interval for real-time monitoring and the frequency of data update need to be determined, so as to improve the accuracy of power system simulation.

[0035] Furthermore, the specific steps of constructing the Bayesian network model are as follows: standardize the collected key data and map the key data after data standardization to the corresponding node positions in the Bayesian network in the form of node variables according to the preset node variables in the power system. Data standardization is used to ensure that the key data input into the power system have the same dimension; obtain the node matching index and construct a directed acyclic graph according to the actual operation principle of the power system and the dependency relationship between the power system structure and the nodes. The node matching index is used to quantify the spatial matching degree between the preset node variables in the power system and the node positions in the Bayesian network. The directed acyclic graph represents the spatial position relationship of the preset node variables in the power system in the Bayesian network model according to the node matching index; set the Bayesian network structure according to the constructed directed acyclic graph and construct a Bayesian network model in combination with the key nodes in the power system. The Bayesian network structure is used to simulate the operation of the power system under the preset scenario. The key nodes include power plants, substations and key data transportation lines.

[0036] In this embodiment, a directed acyclic graph is constructed according to the actual operation principle, structure and dependency relationship between preset node variables of the power system, wherein the preset node variables represent random variables (i.e., the power line parameters corresponding to the power system), and the directed edges represent the dependency relationship between the variables. The nodes are connected through the directed edges through the dependency relationship between the variables to obtain a Bayesian network structure; it should be noted that if the state of a certain device affects the voltage and current of its downstream, then in the directed acyclic graph, the device state node should point to the voltage and current nodes to ensure that the dependency relationship between the nodes in the Bayesian network model forms an acyclic graph; in a preset scenario, the Bayesian network model is used to simulate the operation of the power system. In actual applications, the system response and performance changes can also be observed by setting different input conditions (such as energy supply interruption, equipment failure). Through the simulation results, the stability and reliability of the power system in different scenarios can be evaluated, and corresponding response strategies can be formulated. As the structure and operating conditions of the power system change, the Bayesian network model needs to be regularly updated and maintained, which includes adding new nodes and modifying the parameters in the directed acyclic graph to adapt to new scenarios and requirements, thereby improving the accuracy and reliability of the Bayesian network model.

[0037] Furthermore, the node matching index is obtained by the following method: the key data after data standardization is input into the electrical wiring diagram of the power system to establish an electrical connection relationship diagram between nodes and boundaries, the nodes represent electrical components in the power system, and the changes in the electrical connection distances of the electrical components within a preset time period are monitored in real time to obtain distance measurement indicators, the boundaries represent the electrical connection status between the electrical components, the electrical connection relationship diagram is used to visualize the connection positions of the nodes and boundaries in the power line, and the distance measurement indicators are used to measure the stability of the operating status between the electrical components in the electrical connection diagram; the angle tightness coefficient is obtained according to the angle between the node and the boundary in the electrical connection relationship diagram and the actual needs of the power system, and the functional characteristics and electrical connection angle thresholds between the nodes in the electrical connection relationship diagram are combined with the node matching within a preset time period. The electrical function similarity score is used to obtain the functional similarity. The electrical connection angle threshold represents the angle range between electrical components obtained after weighted summation based on the historical operation data of the power system. The angle tightness coefficient is used to measure the tightness of the electrical connection between the node and the boundary. The functional similarity is used to measure the similarity between the electrical functions of each node within a preset time period. The electrical function data of the nodes within the preset time period are collected, and the fluctuation deviation degree in the electrical function data is adjusted according to the actual simulation requirements of the power system, and the preset functional similarity correction factor is obtained from the database. At the same time, the node matching index is obtained by combining the obtained angle tightness coefficient and functional similarity. The electrical function data includes voltage fluctuation, current fluctuation and power output. The functional similarity correction factor is used to correct the fluctuation deviation degree of the electrical function data during the operation of the power system.

[0038] In this embodiment, monitoring points or sensors are set in the electrical wiring diagram to measure and record the changes of these distances or parameters in real time. The electrical connection distance data monitored in real time are processed according to preset metrics to obtain distance metrics, wherein these metrics usually include average distance, maximum / minimum distance, and the distance metrics represent the ratio of the difference between the maximum distance and the minimum distance to the average distance; in the electrical connection relationship diagram, the angle between each node and the boundary (usually the edge of the power grid or a specific area) is identified and the angle closeness coefficient is obtained in combination with the angle of the electrical component itself. The smaller the angle, the higher the angle closeness coefficient, indicating that the relationship between the node and the boundary is closer; the actual simulation demand of the power system refers to the specific operating state, equipment characteristics and load changes of the power system. The specific operating state includes normal operating state, overload state and short circuit state, wherein the normal operating state refers to the operating state of the power system that meets the power safety specifications, and the overload state refers to the load that the power system bears under the normal operating state. The load received exceeds the rated load capacity, and the duration of the overload state cannot be too long, otherwise it will reduce the stability of the power system; according to the design and operation requirements of the power system, a threshold value of the electrical connection angle is set. Only when the electrical connection angle between two nodes is less than or equal to this threshold, it is considered that there is a direct electrical connection relationship between them. Within a preset time period, the electrical function data of each node (such as power output, voltage fluctuation, and failure rate) is collected. Based on the time sequence of the electrical function data within the preset time period, a weight is determined for each data point. For example, electrical function data that appears earlier is usually considered to be more important and is therefore assigned a higher weight. At the same time, the angle values ​​between electrical components are multiplied by their corresponding weights and summed to obtain the weighted angle sum, that is, the electrical function similarity score between nodes (that is, functional similarity); the node matching index is obtained by weighted summing the obtained angle tightness coefficient and functional similarity, thereby achieving more accurate acquisition of the node matching index.

[0039] Furthermore, the node matching index is calculated by the following formula:

[0040]

[0041] Where r is the number of the electrical component, r = 1, 2, ..., R, R is the total number of electrical components, m is the number of the preset time period, m = 1, 2, ..., M, M is the total number of preset time periods, JIE r.m It represents the node matching index of the rth electrical component in the mth preset time period, e is a natural constant, A r.m represents the distance measurement index of the rth electrical component in the mth preset time period, A0 represents the distance reference measurement index, ΔA represents the distance measurement index reference deviation, Br.m represents the functional similarity of the rth electrical component in the mth preset time period, α represents the functional similarity correction factor, and θ r.m represents the angle tightness coefficient of the rth electrical component in the mth preset time period, and δ represents the electrical connection angle threshold.

[0042] In this embodiment, the distance reference metric is usually obtained by weighted averaging the historical data of the power system in the database that operates well, and the distance metric reference deviation is usually obtained by weighted summing the historical data of the power system in the database that operates well and calculating the difference between the distance reference metric; in θ r.m ≤δ, when And B r.m *When α=1, the node matching index is equal to 2.4. At this time, the spatial matching degree between the preset node variables in the power system and the node positions in the Bayesian network is the highest. When the electrical function data fluctuates greatly during the operation of the power system, the function similarity correction factor increases accordingly. The function similarity correction factor is obtained from a database. The function similarity correction factor in the database can be obtained by fitting the conformity relationship between the function similarity and the node matching during the operation of the power system according to a large amount of historical information data. In a specific embodiment, the relationship between the node matching index and the function similarity is fitted through historical data to obtain a fitting curve. The relationship between the function similarity and the function similarity correction factor is determined according to the fitting curve. The real-time function similarity is substituted into the fitting curve to obtain the corresponding weight factor, thereby improving the accuracy and reliability of the node matching index, thereby improving the accuracy of the power system simulation, and effectively solving the problem of low correlation between the power system simulation accuracy and the real-time operation status of the power system in the prior art.

[0043] Furthermore, the specific steps of real-time monitoring of the operating status of the power system and conducting confidence assessment include: inputting the real-time updated operating status of the power system and the real-time collected key data into the Bayesian network model and monitoring the node status in the Bayesian network model in real time, and updating the conditional probability distribution corresponding to the nodes in the Bayesian network model to obtain the node matching rate. The node matching rate represents the number of matches between the real-time updated key data and the nodes in the Bayesian network model within a preset time period; obtaining the node matching rate according to the updated conditional probability distribution and the degree of matching between the electrical components and the nodes in the Bayesian network model within a preset time period. The node matching rate It indicates the number of matches between electrical components and nodes in the Bayesian network model within a preset time period; the operating status of the power system is evaluated with confidence and the Bayesian network model is overloaded according to the result of the confidence evaluation to obtain a priori probability. At the same time, the obtained node matching rate and node matching rate are combined to obtain the confidence evaluation index. Overload identification is used to identify the load condition of the power line when the electrical components work simultaneously within a preset time period. The prior probability indicates the probability of overload of the electrical components in the power system during operation. The confidence evaluation index is used to measure the reliability evaluation index between the operating status of the power system and the Bayesian network model.

[0044] In this embodiment, the confidence evaluation index is calculated by the following formula:

[0045]

[0046] Wherein, m is the number of the preset time period, m = 1, 2, ..., M, M is the total number of the preset time period, r is the number of the electrical components, r = 1, 2, ..., R, R is the total number of electrical components, y is the number of the key data updated in real time, y = 1, 2, ..., Y, Y is the total number of key data updated in real time, ZH m represents the confidence evaluation index of the Bayesian network model in the mth preset time period, e is a natural constant, K r represents the node matching rate corresponding to the rth electrical component, F y represents the node matching rate corresponding to the yth real-time updated key data, W m It represents the prior probability of the Bayesian network model in the mth preset time period, and W0 represents the reference prior probability.

[0047] Specifically, That is, the calculation formula of the confidence evaluation index is: In the formula, KR r Represents the node matching rate coefficient corresponding to the rth electrical component, FR r Represents the node matching rate coefficient corresponding to the rth electrical component, WR rIt represents the prior probability coefficient corresponding to the rth electrical component. The statistical table of changes in the confidence evaluation index is shown in Table 1:

[0048] Table 1 Statistics of changes in confidence evaluation indicators

[0049]

[0050] It is important to understand that when and ( and ), the confidence evaluation index is equal to 2.6, indicating that the reliability between the operating status of the power system and the Bayesian network model is the highest. The reference prior probability is usually obtained by weighted averaging of historical data, which improves the reliability of the confidence evaluation index and further improves the accuracy of power system simulation, effectively solving the problem of low correlation between the power system simulation accuracy and the real-time operating status of the power system in the prior art.

[0051] Furthermore, the specific process of model training for the Bayesian network model is as follows: according to the actual simulation requirements of the power system, the initial conditional probability distribution of the nodes in the Bayesian network model is set, and at the same time, the load data of the power system in a preset time period is collected in real time and input into the Bayesian network model as training data, and the load data represents the maximum allowable load of the electrical components in the preset time period; the parameters of the Bayesian network model are initialized and the Bayesian network model is trained using the training data to predict the load value of the power system during operation, and at the same time, the model training loss function is defined according to the performance of the power system during operation, and the model training loss function is used to measure the degree of deviation between the load value of the power system during operation and the actual load value; based on the model training loss function and combined with the gradient descent algorithm, the Bayesian network model is iteratively optimized, and the gradient descent algorithm is used to adjust the number of iterations of the model training according to the upward and downward trends of the model performance of the Bayesian network model during the model training process to prevent overfitting, and the iterative optimization is used to re-train the Bayesian network model by changing the network structure and adding regularization terms until the conditional probability distribution of the nodes in the Bayesian network model obtained in real time satisfies the initial conditional probability distribution of the nodes in the Bayesian network model.

[0052] In this embodiment, the prior probability of the node state in the Bayesian network model is collected from historical data, and an initial conditional probability distribution is set for each node. For discrete nodes, this is usually expressed as a conditional probability table, while for continuous nodes, it is usually necessary to initialize a probability density function (such as the parameters of a Gaussian distribution), wherein the prior probability represents the possibility that the node is in a certain state when the electrical components are not running; in actual applications, according to the distribution of the prior probability, the prior probability is weighted and summed within a preset time period to take the expected value, and the load value of the power system during operation is predicted. For the predicted load value, a loss function that measures the prediction accuracy by using the mean square error is selected, and the value of the loss function is obtained according to the predicted load value and the actual load value during the model training process. The parameters of the Bayesian network model are adjusted by the gradient descent algorithm to minimize the loss function. When the actual load value is greater than the predicted load value, the actual load value is reduced by reducing the generator output in the power system. When the actual load value is less than the predicted load value, the actual load value is increased by increasing the capacitance in the power line. After each iterative optimization, the loss function needs to be calculated to evaluate the performance of the Bayesian network model. If the performance of the Bayesian network model has been declining (that is, the value of the loss function has been increasing), it is in an overfitting state. An early stopping condition needs to be set in the Bayesian network model. For example, when the performance of the Bayesian network model has not improved for multiple consecutive iterations, the training is stopped (that is, the iterative optimization is stopped), and the model parameters with the best performance are used as the final parameters of the Bayesian network model to maximize the performance of the Bayesian network model without overfitting, thereby improving the accuracy and reliability of model training.

[0053] Further, such as Figure 2 As shown, it is a flow chart for establishing a simulation power flow circuit provided in an embodiment of the present application. The specific steps for establishing a simulation power flow circuit of an electric power system based on a Bayesian network model after model training are as follows: setting initial parameters of the simulation power flow circuit according to actual simulation requirements of the electric power system, and dynamically adjusting the initial parameters according to the results of model training, wherein the initial parameters include generator output and power line impedance; real-time monitoring of the conditional dependencies between electrical components during the dynamic adjustment of the initial parameters, and obtaining a correlation influence index by visualizing the influence paths between electrical components according to the Bayesian network model, wherein the correlation influence index is used to measure the degree of influence of electrical components on the simulation power flow circuit within a preset time period; simulating the operating state of the Bayesian network model in a preset scenario according to the influence paths and correlation influence indexes between electrical components and in combination with the actual circuit of the electric power system to construct a simulation power flow circuit, and using the output of the Bayesian network model as the input of the simulation power flow circuit to reflect the actual operating state of the electric power system and the power flow distribution of the simulation power flow circuit in real time.

[0054] In this embodiment, according to the load forecast and dispatch plan of the power system, the initial active output (P) and reactive output (Q) of each generator are set, and the operating restrictions of the generator, such as the maximum and minimum output restrictions, are taken into account to ensure that the initial output is within a reasonable range; the resistance (R), inductance (L) and capacitance (C) of the power line are obtained according to the rated voltage, rated current and line length of the power line; these parameters are input into the Bayesian network model as the initial impedance value of the power line, and according to the results of the model training, which parameters have a greater impact on the system performance are analyzed. For the generator output, it can be determined whether the output needs to be increased or decreased according to the adjustment of the load forecast and dispatch plan, as well as the real-time operating status of the system; for the power line impedance, it can be determined whether the impedance value needs to be adjusted according to the actual operating conditions of the line and the simulation results; in the power system, The influence path between electrical components usually refers to the interaction and influence of electrical components within a preset time period. Specifically, starting from a component node and tracing forward along the directed edges, all factors that affect the state of the component (i.e., the influence path) can be found; the power flow distribution usually refers to the voltage distribution, current distribution, and power distribution in the power system, which helps to reflect the conditional dependencies and influence paths between components in the power system in real time; it should be noted that the stepwise approximation method (i.e., fine-tuning the parameters of electrical components) is usually adopted during adjustment to avoid excessive fluctuations in the system. After completing the parameter adjustment, re-run the Bayesian network model to verify whether the adjusted system performance meets the requirements. If not, continue to adjust the parameters. If it does, record the current parameter settings and use them as the initial parameters for the next simulation.

[0055] Specifically, in the Bayesian network model, each electrical component can be represented as a node, the state of the node usually represents the normal operation or fault state of the component, the edge represents the influence relationship between the electrical components, and the probability represents the possibility of the node state occurring under the preset node state, wherein the preset node state usually refers to a certain state assumed by historical data of good operation in the power system (such as the known fault state and the known normal operation state of the electrical component), and then the Bayesian network model is used to analyze the influence paths between the electrical components, which show how the state changes of the electrical components propagate and affect other components; when the distribution pattern of the electrical components in the Bayesian network model is not obvious, the Bayesian network model can be used to quantify the degree of influence of each component on the power flow distribution within a preset time period, that is, the correlation influence index, especially focusing on the influence of components with a higher correlation influence index on the power flow distribution, which helps to better understand the changes in the operating state of the power system during the simulation process and improve the stability and reliability of the simulated power flow circuit.

[0056] Furthermore, the specific steps for obtaining the correlation influence index are as follows: collecting the starting and ending values ​​of the initial parameters of the power system during operation under a preset scenario, and monitoring the numerical changes of the electrical components in the process of dynamically adjusting the initial parameters in real time to obtain the electrical influence coefficient and obtain the preset initial parameter correction factor from the preset database. The electrical influence coefficient is used to measure the influence of the electrical components on the initial parameters within a preset time period; real-time monitoring of the influence of the electrical components of the simulated power flow circuit on the power flow distribution path under the preset scenario to obtain the distribution deviation rate and dynamically adjust the power flow distribution path according to the preset path of the power flow distribution, and obtain the preset dynamic adjustment factor from the preset database. The distribution deviation rate is used to measure the deviation between the power flow distribution path and the preset path of the power flow distribution; combining the obtained electrical influence coefficient and the initial parameter correction factor to obtain the correlation influence index, and the correlation influence index is calculated by the following formula:

[0057]

[0058] In the formula, r is the number of the electrical component, r = 1, 2, ..., R, R is the total number of electrical components, n is the number of the preset scene, n = 1, 2, ..., N, N is the total number of preset scenes, GU r.n It represents the correlation impact index of the rth electrical component in the nth preset scenario, e is a natural constant, Indicates the initial parameter end value of the rth electrical component in the nth preset scenario, represents the initial parameter starting value of the rth electrical component in the nth preset scenario, ΔP represents the initial parameter reference deviation, D r.n It represents the electrical influence coefficient of the rth electrical component in the nth preset scenario, Q r.n It represents the distribution deviation rate of the rth electrical component in the nth preset scenario, Q0 represents the reference distribution deviation rate, γ represents the dynamic adjustment factor, and β represents the initial parameter correction factor.

[0059] In this embodiment, within a preset time period, an electrical component is randomly selected as a reference point to observe the impact of its parameter changes on other electrical components, and the location of the electrical component whose impact on the affected point needs to be evaluated is determined. Then, the electrical parameters (such as voltage and current) of the reference point and the affected point are monitored in real time, and the change in the parameters of the affected point when the reference point parameters change is counted and divided by the total parameter change to obtain the electrical influence coefficient; in practical applications, the distribution deviation rate is usually obtained by comparing the simulation results of the power system under a preset scenario with the power flow distribution path under the benchmark situation. For example, the voltage distribution deviation rate represents the ratio of the absolute value of the difference between the actual voltage and the rated voltage to the rated voltage, and the power distribution deviation rate represents the ratio of the absolute value of the difference between the actual power and the rated power to the rated power; In the formula, AI r.n Indicates the electrical impact index of the rth electrical component in the nth preset scenario, BI r.n It represents the distribution deviation rate index of the rth electrical component in the nth preset scenario. The reference distribution deviation rate is usually obtained by weighted averaging the historical data. The initial parameter reference deviation is usually obtained by weighted summing the operating parameters of the known electrical components and subtracting the parameters of the actual operating process. In a specific embodiment, a fitting curve is obtained through the relationship between historical data, the correlation influence index, the electrical influence coefficient and the initial parameters. The relationship between the initial parameters and the initial parameter correction factor is determined according to the fitting curve, and the real-time electrical influence coefficient is substituted into the fitting curve to obtain the corresponding correction factor.

[0060] Specifically, Figure 3 As shown, it is a three-dimensional statistical analysis diagram of the correlation influence index provided in the embodiment of the present application. It can be seen from the figure that the distribution deviation rate index has a smaller influence on the correlation influence index, while the electrical influence index has a larger influence on the correlation influence index. When the correlation influence index is equal to 1.5, it indicates that the electrical components have the greatest influence on the simulated power flow circuit within the preset time period. It should be understood that, in a specific embodiment, the dynamic adjustment factor in the database can expand the influence of the distribution relative deviation rate on the correlation influence index, and the reference distribution deviation rate can be adjusted according to the required sensitivity and form a mapping table with the dynamic adjustment factor. The real-time adjusted reference distribution deviation rate is input into the mapping table to obtain the corresponding dynamic adjustment factor; wherein, the distribution relative deviation rate represents the ratio of the distribution deviation rate to the reference distribution deviation rate. When the distribution relative deviation rate increases, the corresponding dynamic adjustment factor will also increase accordingly, thereby achieving more accurate acquisition of the correlation influence index, and then achieving improved accuracy of power system simulation, and effectively solving the problem of low correlation between power system simulation accuracy and real-time operating status of the power system in the prior art.

[0061] Furthermore, after real-time monitoring of the changing trends of the voltage and current values ​​on the power lines during the simulation process of the simulated power flow circuit, the method also includes adjusting the circuit parameters to be adjusted in the simulated power flow circuit. The specific adjustment steps include: real-time acquisition of simulation data of the simulated power flow circuit during the power system simulation process and comparison with the rated data input into the simulated power flow circuit to determine the circuit parameters to be adjusted, the simulation data including voltage data and current data, the rated data including rated voltage data and rated current data, and the circuit parameters to be adjusted are used to reflect the distribution of the power flow in the power system within a preset time period; according to the results of the data comparison, the changing trends of the simulation data are identified and the circuit parameters to be adjusted in the simulated power flow circuit are adjusted, and at the same time, the adjusted circuit parameters are input into the simulated power flow circuit and the power system simulation is re-performed until the actual simulation requirements of the power system are met.

[0062] In this embodiment, power system simulation software is usually used to simulate the operation process of the power system in a simulated power flow circuit; the circuit parameters to be adjusted usually refer to circuit parameters that affect the power flow distribution, usually including resistance, capacitance and inductance; the simulation data acquired in real time is compared with the rated data input into the simulated power flow circuit, and by comparing the difference between the two, the areas or nodes that deviate from the rated values ​​in the simulation results can be identified. These areas or nodes usually require parameter adjustment. The specific process is: when the simulated voltage data in the power line is greater than the rated voltage data, the resistance or inductance of the power line is increased to reduce the simulated voltage data; when the simulated current data in the power line is greater than the rated When adjusting the current data, increase the capacitance of the power line or reduce the load to reduce the simulated current data; input the adjusted circuit parameters into the simulated power flow circuit to re-simulate the power system and compare it with the rated data again; the change trend of the simulation data usually refers to the direction and speed of increase or decrease of the simulation data over time or other variables, usually including the increase or decrease of electrical parameters in the power system, such as voltage, current, and power factor. The change trend of these parameters can be judged by fitting the curve of the simulation data points with the least squares method. When the slope is positive, it means that the data is on an upward trend; when the slope is negative, it means that the data is on a downward trend, thereby improving the accuracy and reliability of the power system simulation.

[0063] like Figure 4As shown, it is a structural schematic diagram of a power system simulation system based on a Bayesian network provided in an embodiment of the present application. The power system simulation system based on a Bayesian network provided in an embodiment of the present application includes: a Bayesian network model construction module, a confidence assessment module, a model training module and a simulation flow circuit establishment module; wherein the Bayesian network model construction module is used to collect key data of the power system within a preset time period and construct a Bayesian network model, the key data including supply voltage data, load change data and transmission circuit data, the node variables are used to numerically describe the dependency relationship between voltage, current and power of the power system during operation, and the Bayesian network model is used to predict the load change of the power system within a preset time period; the confidence assessment module is used to monitor the operating status of the power system in real time and perform confidence assessment according to the conditional probability distribution of the key data in the Bayesian network model, the conditional The probability distribution is used to describe the probability of a circuit failure in the power system under preset voltage conditions, and the confidence assessment is used to quantify the credibility of the conditional probability distribution results; the model training module is used to obtain the power grid data of the power system during operation within a preset time period and train the Bayesian network model according to the results of the confidence assessment. The power grid data represents the real-time operating status of the power system under a preset load. The model training means inputting the output of the generator and the load distribution of the power system into the Bayesian network model and training it according to the results of the confidence assessment; the simulation power flow circuit establishment module is used to establish a simulation power flow circuit of the power system based on the Bayesian network model after model training, and at the same time monitor the changing trends of the voltage and current values ​​on the power lines of the simulation power flow circuit in real time during the simulation process. The simulation power flow circuit is used to visualize the influence of electrical components in the simulation power system on the power flow distribution.

[0064] In this embodiment, in the Bayesian network, the conditional probability distribution describes the dependency between variables. For the power system, the conditional probability distribution may be used to describe the probability of a circuit failure under given voltage conditions. For example, when the voltage of a power line is monitored to be out of the normal range, the conditional probability distribution can give the probability of failure of the line; the confidence assessment module can help operators better understand the operating status of the system. When the system fails or an abnormal situation occurs, the module can quickly give the probability and credibility of the failure, helping the operator to respond in time and take corresponding maintenance or control measures. In addition, the confidence assessment module can also be combined with the simulation flow circuit establishment module to verify the accuracy of the confidence assessment by simulating the operation of the power system; in practical applications, the output of the generator and the transformation ratio of the transformer in the electrical components are usually adjusted to observe the effect of these changes on the power flow distribution of the power system in the simulation flow circuit, and a bar chart is used to represent the power injection and outflow on the power line. At the same time, the simulation flow circuit will re-input these changes into the simulation process of the power system and recalculate the power flow of the power system, thereby improving the stability and safety of the power system.

[0065] To summarize, the embodiment of the present application collects key data of the power system within a preset time period and constructs a Bayesian network model, then monitors the operating status of the power system in real time and performs confidence assessment, and then trains the Bayesian network model according to the results of the confidence assessment. Finally, based on the trained Bayesian network model, a simulated power flow circuit of the power system is established, and at the same time, the changing trends of the voltage and current values ​​on the power lines in the simulated power flow circuit during the simulation process are monitored in real time, thereby achieving a more accurate establishment of the simulated power flow circuit, and then achieving an improvement in the accuracy of power system simulation, effectively solving the problem of low correlation between the power system simulation accuracy and the real-time operating status of the power system in the prior art.

[0066] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0067] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0068] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0069] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0070] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0071] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A power system simulation method based on Bayesian network, characterized in that: The following steps are involved: S1, collect key data of the power system within a preset time period and build a Bayesian network model; S2, monitors the operating status of the power system in real time and performs confidence assessment based on the conditional probability distribution of key data in the Bayesian network model; S3, obtaining the grid data of the power system during operation within a preset time period and performing model training on the Bayesian network model according to the confidence evaluation result; S4, based on the Bayesian network model after model training, establish a simulated power flow circuit of the power system, and monitor in real time the changing trend of the voltage and current values ​​on the power line during the simulation process of the simulated power flow circuit; The specific steps of real-time monitoring of the operating status of the power system and performing confidence assessment include: Inputting the real-time updated power system operation status and the real-time collected key data into the Bayesian network model and monitoring the node status in the Bayesian network model in real time, and updating the conditional probability distribution corresponding to the node in the Bayesian network model accordingly to obtain the node matching rate, wherein the node matching rate represents the number of matches between the real-time updated key data and the nodes in the Bayesian network model within a preset time period; Obtaining a node matching rate according to the updated conditional probability distribution and the matching degree between the electrical components and the nodes in the Bayesian network model within a preset time period, wherein the node matching rate represents the number of matches between the electrical components and the nodes in the Bayesian network model within the preset time period; Conduct confidence assessment on the operating status of the power system and identify overloads on the Bayesian network model based on the confidence assessment results to obtain a priori probability. Combine the obtained node matching rate and node coincidence rate to obtain a confidence assessment index. The specific steps of establishing the simulation power flow circuit of the power system based on the Bayesian network model after model training are as follows: The initial parameters of the simulation power flow circuit are set according to the actual simulation requirements of the power system, and the initial parameters are dynamically adjusted according to the results of model training. The initial parameters include generator output and power line impedance. Real-time monitoring of the conditional dependencies between electrical components during the dynamic adjustment of initial parameters, and visualization of the impact paths between electrical components based on the Bayesian network model to obtain the correlation impact index, which is used to measure the degree of influence of electrical components on the simulated power flow circuit within a preset time period; According to the influence path and correlation influence index between electrical components and combined with the actual circuit simulation of the power system, the operating status of the Bayesian network model in the preset scenario is constructed to construct a simulated power flow circuit. At the same time, the output of the Bayesian network model is used as the input of the simulated power flow circuit to reflect the actual operating status of the power system and the power flow distribution of the simulated power flow circuit in real time.

2. A power system simulation method based on Bayesian network as claimed in claim 1, characterized in that: The specific construction steps of the Bayesian network model are: Standardize the collected key data and map the standardized key data to the corresponding node positions in the Bayesian network in the form of node variables according to the preset node variables in the power system; According to the actual operation principle of the power system and the dependency relationship between the power system structure and nodes, a node matching index is obtained and a directed acyclic graph is constructed. The node matching index is used to quantify the spatial matching degree between the preset node variables in the power system and the node positions in the Bayesian network; The Bayesian network structure is set according to the constructed directed acyclic graph and the Bayesian network model is constructed in combination with the key nodes in the power system.

3. A power system simulation method based on Bayesian network as claimed in claim 2, characterized in that: The node matching index is obtained by the following method: The standardized key data is input into the electrical wiring diagram of the power system to establish an electrical connection relationship diagram between nodes and boundaries, and at the same time, the changes in the electrical connection distance of electrical components within a preset time period are monitored in real time to obtain distance measurement indicators; The angle closeness coefficient is obtained according to the angle between the node and the boundary in the electrical connection relationship diagram and the actual demand of the power system, and the functional similarity is obtained according to the functional characteristics and electrical connection angle threshold between the nodes in the electrical connection relationship diagram combined with the electrical function similarity score corresponding to the node in a preset time period, wherein the electrical connection angle threshold represents the angle range between the electrical components obtained after weighted summation based on the historical operation data of the power system; The electrical function data of the nodes within a preset time period are collected, the fluctuation deviation degree in the electrical function data is adjusted according to the actual simulation requirements of the power system, and a preset function similarity correction factor is obtained from the database. At the same time, the node matching index is obtained by combining the obtained angle closeness coefficient and function similarity. The electrical function data includes voltage fluctuation, current fluctuation and power output, and the function similarity correction factor is used to correct the fluctuation deviation degree of the electrical function data during the operation of the power system.

4. A power system simulation method based on Bayesian network as claimed in claim 3, characterized in that: The node matching index is calculated by the following formula: Where r is the number of the electrical component, r = 1, 2, ..., R, R is the total number of electrical components, m is the number of the preset time period, m = 1, 2, ..., M, M is the total number of preset time periods, JIE r.m It represents the node matching index of the rth electrical component in the mth preset time period, e is a natural constant, A r.m represents the distance measurement index of the rth electrical component in the mth preset time period, A0 represents the distance reference measurement index, ΔA represents the distance measurement index reference deviation, B r.m represents the functional similarity of the rth electrical component in the mth preset time period, α represents the functional similarity correction factor, and θ r.m represents the angle tightness coefficient of the rth electrical component in the mth preset time period, and δ represents the electrical connection angle threshold.

5. A method for simulating a power system based on a Bayesian network as claimed in claim 1, characterized in that: The specific process of training the Bayesian network model is as follows: The initial condition probability distribution of the nodes in the Bayesian network model is set according to the actual simulation requirements of the power system, and the load data of the power system in a preset time period is collected in real time and input into the Bayesian network model as training data; Initialize the parameters of the Bayesian network model and use the training data to train the Bayesian network model to predict the load value of the power system during operation, and define the model training loss function according to the performance of the power system during operation; The Bayesian network model is iteratively optimized based on the model training loss function and combined with the gradient descent algorithm.

6. A method for simulating a power system based on a Bayesian network as claimed in claim 1, characterized in that: The specific steps for obtaining the relevance impact index are: Collect the starting and ending values ​​of the initial parameters of the power system during operation under the preset scenario, and monitor the value changes of the electrical components in the process of dynamic adjustment of the initial parameters in real time to obtain the electrical influence coefficient and obtain the preset initial parameter correction factor from the preset database; Real-time monitoring of the influence of electrical components on the power flow distribution path of the simulated power flow circuit in a preset scenario to obtain the distribution deviation rate and dynamically adjust the power flow distribution path according to the preset path of the power flow distribution, and at the same time obtain the preset dynamic adjustment factor from the preset database; The obtained electrical influence coefficient and the initial parameter correction factor are combined to obtain the correlation influence index, which is calculated by the following formula: In the formula, r is the number of the electrical component, r = 1, 2, ..., R, R is the total number of electrical components, n is the number of the preset scene, n = 1, 2, ..., N, N is the total number of preset scenes, GU r.n It represents the correlation impact index of the rth electrical component in the nth preset scenario, e is a natural constant, Indicates the initial parameter end value of the rth electrical component in the nth preset scenario, represents the initial parameter starting value of the rth electrical component in the nth preset scenario, ΔP represents the initial parameter reference deviation, D r.n It represents the electrical influence coefficient of the rth electrical component in the nth preset scenario, Q r.n It represents the distribution deviation rate of the rth electrical component in the nth preset scenario, Q0 represents the reference distribution deviation rate, γ represents the dynamic adjustment factor, and β represents the initial parameter correction factor.

7. A method for simulating a power system based on a Bayesian network as claimed in claim 1, characterized in that: The real-time monitoring of the change trend of the voltage and current values ​​on the power line during the simulation process of the simulated power flow circuit also includes adjusting the circuit parameters to be adjusted in the simulated power flow circuit. The specific adjustment steps include: Acquire simulation data of the simulation power flow circuit in the process of power system simulation in real time and compare it with the rated data input into the simulation power flow circuit to determine the circuit parameters to be adjusted, wherein the simulation data includes voltage data and current data, and the rated data includes rated voltage data and rated current data; According to the results of data comparison, the changing trend of simulation data is identified and the circuit parameters to be adjusted in the simulation power flow circuit are adjusted. At the same time, the adjusted circuit parameters are input into the simulation power flow circuit and the power system simulation is re-performed until the actual simulation requirements of the power system are met.

8. A power system simulation system based on Bayesian network, characterized in that: A Bayesian network-based power system simulation method for executing any one of claims 1 to 7, comprising: a Bayesian network model building module, a confidence assessment module, a model training module and a simulation power flow circuit establishment module; The Bayesian network model building module is used to collect key data of the power system within a preset time period and build a Bayesian network model; The confidence assessment module is used to monitor the operating status of the power system in real time and perform confidence assessment according to the conditional probability distribution of key data in the Bayesian network model; The model training module is used to obtain the power grid data of the power system during operation within a preset time period and perform model training on the Bayesian network model according to the result of the confidence evaluation; The simulation power flow circuit establishment module is used to establish the simulation power flow circuit of the power system based on the Bayesian network model after model training, and at the same time monitor the change trend of the voltage and current values ​​on the power line of the simulation power flow circuit in real time during the simulation process.

Images