A power system risk warning method, device, storage medium and product

Abstract

The application discloses a power system risk warning method and device, a storage medium and a product, and relates to the technical field of power system risk warning. The method comprises the following steps: constructing data samples at each moment; inputting the active power of a generator at the current moment, the reactive power of the generator at the current moment, the active power of a load at the current moment, the reactive power of the load at the current moment, the topological state at the current moment and meteorological data at the next moment into a preset topological prediction model, calculating the topological state at the next moment, inputting the data samples at a plurality of moments in history, the topological state at the next moment and the meteorological data at the next moment into a preset operation state prediction model, and obtaining the bus voltage at the next moment and the load rate at the next moment; for the next moment, calculating the risk value of the power system according to the topological state, the bus voltage and the load rate; and performing a warning operation on the power system according to the risk value. The accuracy of power system risk prediction is improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of power system reliability analysis, and in particular to a power system risk alarm method, device, storage medium and product. Background Technology

[0002] As power systems continue to expand and become more complex, they are affected by various factors during operation, such as load fluctuations, equipment failures, and weather changes. These factors can lead to operational anomalies or even malfunctions, impacting the reliability and economy of power supply. Therefore, accurately predicting the state of the power system is one of the key factors in improving its security.

[0003] Traditional methods for judging the operating status of power systems mainly rely on rule-based system monitoring and manual analysis. These methods have low adaptability and analytical capabilities for complex factors such as load fluctuations, equipment failures, and weather changes, resulting in low accuracy in predicting power system risks. Summary of the Invention

[0004] This invention provides a power system risk alarm method, device, storage medium, and product to improve the accuracy of power system risk prediction.

[0005] In a first aspect, embodiments of the present invention provide a power system risk alarm method, comprising:

[0006] Data samples are constructed at various times; each data sample includes bus voltage, generator active power, generator reactive power, load active power, load reactive power, power system topology, load rate of transmission lines and transformers, and meteorological data.

[0007] The current generator active power, the current generator reactive power, the current load active power, the current load reactive power, the current topology state, and the meteorological data for the next time are input into a preset topology prediction model to calculate the topology state for the next time.

[0008] The data samples from multiple historical moments, the topology state at the next moment, and the meteorological data at the next moment are input into a preset operation state prediction model to calculate the bus voltage and the load rate at the next moment.

[0009] For the next moment, calculate the risk value of the power system based on the topology state, the bus voltage, and the load rate;

[0010] An alarm operation is performed on the power system based on the risk value.

[0011] Secondly, embodiments of the present invention provide a power system risk alarm device, comprising:

[0012] The data sample construction module is used to construct data samples at various times; each of the data samples includes bus voltage, generator active power, generator reactive power, load active power, load reactive power, power system topology, load rate of transmission lines and transformers, and meteorological data.

[0013] The topology prediction module is used to input the generator active power, generator reactive power, load active power, load reactive power, topology state, and meteorological data at the next moment into a preset topology prediction model to calculate the topology state at the next moment.

[0014] The motion state prediction module is used to input the data samples from multiple historical moments, the topology state at the next moment, and the meteorological data at the next moment into a preset operation state prediction model to calculate the bus voltage and the load rate at the next moment.

[0015] The risk value calculation module is used to calculate the risk value of the power system for the next time step based on the topology state, the bus voltage, and the load rate.

[0016] The power system alarm module is used to perform alarm operations on the power system based on the risk value.

[0017] Thirdly, embodiments of the present invention also provide a computer device, the computer device comprising:

[0018] One or more processors;

[0019] Storage device for storing one or more programs;

[0020] When the one or more programs are executed by the one or more processors, the one or more processors implement the power system risk alarm method as provided in the first aspect of the present invention.

[0021] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the power system risk alarm method as provided in the first aspect of the present invention.

[0022] Fifthly, embodiments of the present invention also provide a computer program product, the computer program product including a computer program, which, when executed by a processor, implements the power system risk alarm method provided in the first aspect of the present invention.

[0023] In this embodiment of the invention, data samples are constructed at various times. Each data sample includes bus voltage, generator active power, generator reactive power, load active power, load reactive power, power system topology, load rate of transmission lines and transformers, and meteorological data. The current generator active power, generator reactive power, load active power, load reactive power, topology, and meteorological data for the next time are input into a preset topology prediction model to calculate the topology for the next time. Data samples from multiple historical times, the topology for the next time, and the meteorological data for the next time are input into a preset operating state prediction model to obtain the bus voltage and load rate for the next time. For the next time, the risk value of the power system is calculated based on the topology, bus voltage, and load rate. Based on the risk value, an alarm operation is performed on the power system. By predicting the topology, operating state, and risk value, potential faults can be detected and warned in advance, thereby reducing the risk of power outages and accidents. Meanwhile, the integration of meteorological data enhances the power system's ability to respond to emergencies. By comprehensively considering multiple data sources and sequentially using topology prediction models and operational status prediction models to predict power system risks, the correlation between various data sources is strengthened, and the accuracy of power system risk prediction is improved. Attached Figure Description

[0024] Figure 1 A flowchart of a power system risk alarm method provided in Embodiment 1 of the present invention;

[0025] Figure 2 This is a schematic diagram of the network structure of a topology prediction model provided in Embodiment 1 of the present invention;

[0026] Figure 3 This is a schematic diagram of the structure of a running state prediction model provided in Embodiment 1 of the present invention;

[0027] Figure 4 This is a schematic diagram of a power system risk alarm provided in Embodiment 1 of the present invention;

[0028] Figure 5 This is a structural block diagram of a power system risk alarm device provided in Embodiment 2 of the present invention;

[0029] Figure 6 This is a schematic diagram of the structure of a computer device provided in Embodiment 3 of the present invention. Detailed Implementation

[0030] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be used interchangeably where appropriate so that the embodiments of the invention described herein can cover implementations in sequences other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0032] Example 1

[0033] See Figure 1 The diagram illustrates a flowchart of a power system risk alarm method according to Embodiment 1 of the present invention. This method can be executed by a power system risk alarm device, which can be implemented in hardware and / or software and can be configured in a computer device. Figure 1 As shown, the method includes:

[0034] Step 101: Construct data samples at each time point.

[0035] In this embodiment, data samples of the power system at various times are collected and accumulated. These times include the current time, future time, and historical time. These data samples include bus voltage, generator active power, generator reactive power, load active power, load reactive power, power system topology, transmission line and transformer load rates, and meteorological data. These data samples provide the foundation for subsequent power system forecasting and risk assessment. By using data samples from multiple times, the dynamic changes and potential risks of the power system can be better reflected, ensuring that the state of the power system at each time point is comprehensively recorded and accurately analyzed. This provides multi-dimensional reference information for risk assessment, thereby effectively predicting the future state of the power system and identifying potential operational risks in advance.

[0036] For example, data samples are collected from monitoring power systems and data acquisition databases. Meteorological data from the meteorological information center includes weather forecasts, typhoon forecasts, and real-time typhoon information. The temporal resolutions for normal weather and typhoons are 48-hour and 24-hour hourly forecasts, respectively. The spatial resolution of weather forecasts is 5 km. Specifically, the temporal resolution of wind-related forecast data ranges from 5 minutes to 3 hours. Since risk assessment of power systems requires a large number of data samples, more virtual data samples can be generated. A Gaussian distribution is chosen to generate a large number of virtual simulation data samples.

[0037] Step 102: Input the current generator active power, current generator reactive power, current load active power, current load reactive power, current topology state, and next time meteorological data into the preset topology prediction model to calculate the next time topology state.

[0038] Topology refers to the connection relationships and operating states between various devices (such as generators, transformers, and lines) in a power system, including the switching status and fault conditions of the devices. Topology determines the path of power flow, system stability, and the response to faults. Changes in topology usually imply adjustments to the power system structure, possibly due to equipment failures, dispatching, or other factors, affecting the safety and reliability of the power system. The topology of this invention includes both the topology of the power system and the probability of that topology.

[0039] In this embodiment, key variables such as the current generator active power, current generator reactive power, current load active power, current load reactive power, and the power system topology are input into a pre-set topology prediction model. Combined with meteorological data for the next time step, calculations are performed to pre-identify potential topology states, providing foundational data for further risk assessment. This improves the foresight and responsiveness of the power system, thereby optimizing its operational stability and security.

[0040] Topology prediction models employ stochastic neural networks, which can be Bayesian neural networks. Bayesian neural networks combine the advantages of Bayesian inference and neural networks, enabling effective learning and prediction even under conditions of high uncertainty. By introducing a probability distribution to represent the uncertainty of the weights, Bayesian neural networks become more robust to noisy data or limited samples, making them suitable for tasks requiring high reliability and risk assessment.

[0041] An example is the improvement from artificial neural networks to Bayesian neural networks. Essentially, artificial neural networks can describe arbitrarily complex mappings y = f(x) between inputs and outputs. Traditional artificial neural networks consist of an input layer l0 and a series of hidden layers l... i i = 0, 1, ..., n-1, with output layer l n .

[0042] l0 = x;

[0043] l i =s i (w i *l i-1 +b i );

[0044] y = l n ;

[0045] In the formula, w i b represents the weight. i Indicates bias, s i Let x represent the activation function, l0 represent the input, and l represent the input layer. i y represents the hidden layer, l represents the output, and y represents the output. n Indicates the output layer.

[0046] Because the parameters of an artificial neural network are fixed (i.e., the weights W) i and bias b i In a probabilistic neural network (PLN), each parameter has a fixed value, so given the same input, the output will not change. However, weather-related power system failures are probabilistic random events, and artificial neural networks lack the ability to simulate such problems. Therefore, stochastic neural networks are introduced into the modeling of probabilistic random problems. Stochastic neural networks (such as Bayesian neural networks) are artificial neural networks that incorporate random components into the network. Stochastic neural networks can simulate multiple possible models by giving the network random parameters. This process can be summarized as follows:

[0047] θ~p(θ); y=f θ (x)+∈

[0048] In the above formula, θ~p(θ) indicates that the network parameter θ follows a probability distribution p(θ), ∈ represents random noise, i.e., random perturbation, y represents the output of the random neural network, and f θ (x) represents the output data processed by a random neural network. A Bayesian neural network can then be defined as any random neural network trained using Bayesian inference.

[0049] In one embodiment of the present invention, step 102 may include the following steps:

[0050] Step 1021: Set the posterior distribution for the preset topology prediction model.

[0051] In this embodiment, the purpose of setting a posterior distribution is to model the uncertainty of the power system's topological state using Bayesian inference methods, thereby obtaining the probability distribution of the power system's future topological state. The posterior distribution provides a framework that, based on known sample data at the current moment, and combined with prior knowledge of the power system (such as historical topological states and meteorological data), can calculate the probability distribution of the power system's topological state at the next moment. This is not merely about obtaining a prediction value, but rather about understanding the likelihood and risk of topological changes from a probabilistic perspective. Therefore, setting a posterior distribution helps the model consider multiple possible power system topological states during prediction and introduces uncertainty and bias during inference, further improving the reliability and robustness of the prediction. The topology model includes an input layer, a hidden layer, and an output layer.

[0052] In one embodiment of the present invention, step 1021 includes the following steps:

[0053] Step 10211: Construct the posterior distribution function for the pre-set topological prediction model based on Bayes' theorem.

[0054] In this embodiment, the core purpose of constructing the posterior distribution function is to model the uncertainty of the power system topology using Bayes' theorem, enabling more reliable predictions and risk assessments in uncertain environments. Bayes' theorem provides a mathematical framework that combines prior knowledge (i.e., previous power system topology states and assumptions) with new observational data (such as the current power system topology state). Specifically, by combining the prior distribution, likelihood function, and marginal likelihood, Bayes' theorem can update and obtain a more accurate posterior distribution. By constructing the posterior distribution function for the topology prediction model, the probability of each possible topology state of the power system occurring at the next time step can be quantified, which is typically important for further risk assessment, alarm, and dispatch decisions.

[0055] For example, the posterior distribution function is expressed as:

[0056]

[0057] In the formula, p(θ|D) represents the posterior distribution function, θ represents the network parameters of the topology prediction model, D represents the data sample, p(θ) represents the prior distribution, p(D|θ) represents the likelihood function, and p(D) represents the marginal likelihood. This represents the i-th data sample. This represents the i-th topological state. Indicates in Given θ, the probability of the i-th topological state occurring, where N represents the number of data samples. θ includes a local random parameter β and a global deterministic parameter α.

[0058] The Bayesian posterior distribution of complex models (such as artificial neural networks) is a high-dimensional, highly non-convex probability distribution. This complexity makes computing and sampling the Bayesian posterior using standard methods a challenging problem, especially given the computational evidence. It is difficult, and to solve this problem, distributed reasoning is introduced.

[0059] Step 10212: Perform distribution inference based on the posterior distribution function to obtain the approximate posterior of the posterior distribution as the topological model.

[0060] In this embodiment, the purpose of performing distribution inference during Bayesian inference is to approximate the specific form of the posterior distribution using certain mathematical methods, so as to further infer and predict. Since the precise calculation of the posterior distribution is usually very complex, especially for high-dimensional systems (such as power system topology prediction models), directly calculating the posterior distribution may be impractical. Therefore, it is necessary to simplify the problem through approximate inference methods to obtain a reasonable approximate posterior distribution.

[0061] Distributed inference is an optimization algorithm that approximates a complex posterior distribution by introducing a simple variational distribution. It then optimizes the approximation by minimizing the difference (divergence) between the variational distribution and the true posterior distribution. In this way, a sufficiently accurate approximate posterior distribution can be obtained while maintaining computational feasibility, thus supporting topological state prediction and future risk assessment.

[0062] In one embodiment of the present invention, step 10212 may include the following steps:

[0063] Step 2121: In the process of performing distribution inference on the posterior distribution function, a variational distribution is set.

[0064] In this embodiment, the variational distribution is set to approximate a complex posterior distribution by selecting a simplified and easily tractable distribution. The key to distributional inference is introducing a simple hypothetical distribution (variational distribution), typically a family of distributions (such as a Gaussian distribution) to approximate the true posterior distribution. This simplified variational distribution can be computed in a shorter time and provides an acceptable approximation. By setting the variational distribution, optimization algorithms can be used to find the optimal variational distribution without complex integration, ultimately approximating the true posterior distribution. The choice of variational distribution directly affects the accuracy and computational efficiency of subsequent inference.

[0065] Step 2122: Calculate the divergence between the variational distribution and the posterior distribution function to obtain the lower bound of evidence.

[0066] In this embodiment, when a variational distribution is set, it is necessary to evaluate the difference between this variational distribution and the true posterior distribution. The divergence between the variational distribution and the posterior distribution is calculated, and through this calculation, the Evidence Lower Bound (ELBO) is obtained, which is an important indicator for measuring how closely the posterior distribution approximates the true posterior distribution.

[0067] Divergence is a mathematical tool for quantifying the difference between two probability distributions. Common divergence measures include the Kullback-Leibler (KL) divergence. A smaller divergence indicates that the variational distribution is closer to the true posterior distribution, leading to higher accuracy in distribution inference. In distribution inference, the goal is to minimize divergence to optimize the accuracy of the variational distribution.

[0068] For example, distributed reasoning transforms a reasoning problem into an optimization problem. Distributed reasoning introduces a set of distributions. This is called the variational distribution. These are the parameters of this variational distribution, and then... Optimization is performed to find the distribution family members that best approximate the posterior of interest. A commonly used proximity measure is the Kullback-Leibler divergence. The KL divergence is used to obtain the lower bound of evidence, and an optimization approximation method is employed to approximate the posterior distribution, which is difficult to solve directly. The lower bound of evidence is expressed as:

[0069]

[0070] In the formula, Let θ represent the divergence between the variational distribution and the posterior distribution function, and let θ represent the network parameters of the topology prediction model. Let denote the variational distribution, p(θ|D) denote the posterior distribution function, p(D,θ) denote the joint distribution, and p(D) denote the marginal likelihood. express Expected value, IE q [logp(D,θ)] represents the expected value of p(D,θ). Indicates the lower bound of the evidence; const indicates a constant.

[0071] logp(D) is replaced by a constant const because logp(D) does not depend on θ, which represents the network parameters of the topology prediction model, including the local random parameter β and the global deterministic parameter α.

[0072] Step 2123: Optimize the lower bound of evidence based on the loss function and Bayesian point estimation to obtain the approximate posterior distribution as the topological model.

[0073] In this embodiment, a lower bound of evidence is obtained by setting a variational distribution and calculating its divergence. This lower bound is then optimized to obtain a final approximate posterior of the variational distribution, enabling a more accurate simulation of changes in the power system's topology. Optimizing the lower bound typically involves using a loss function and Bayesian point estimation. These tools help adjust the variational distribution to approximate the true posterior distribution as closely as possible. The loss function measures the difference between the variational distribution and the true distribution. Through Bayesian point estimation, the power system can select a specific point estimate (such as maximum a posteriori estimation) rather than the complete distribution, thus providing a more reasonable estimation result even with incomplete information. By optimizing the loss function, the power system can obtain an efficient and reliable approximate posterior distribution in practical applications. This approximate posterior provides a core basis for power system topology prediction, risk assessment, and alarm systems.

[0074] For example, by calculating the KL divergence mentioned above, a lower bound on the evidence can be defined. To obtain an approximate posterior of the posterior distribution of the topological model, we need to minimize KL, which is equivalent to maximizing

[0075] According to Bayesian decomposition Decomposed into parts related to β The part related to α The maximization is obtained by optimizing the local random parameter β and the global deterministic parameter α. The following 1-2 are the parts related to solving β. The part related to α

[0076] 1. Solve for the part related to β.

[0077] Solve for the dependencies of β among the parameters in the parameter space:

[0078] p(β n |y n ,β -n ,α)=h(β n )exp{η(y n ,β -n ,α) T T(β n )-a(η(y n ,β -n ,α)};

[0079] In the formula, p(β) n |y n ,β -n ,α) represents the condition where y is given n ,β -nIn the case of α, β n The conditional probability distribution, β n Let y represent the nth local random parameter. n Let β represent the nth topological state. -n Indicates the difference from β n Local random parameters other than β, h(β) n ) represents β n The underlying measure, T(β) n ) represents β n A sufficient statistic, η(y) n ,β -n ,α) represents the natural parameter, a(η(y) n ,β -n ,α) denotes the cumulative function.

[0080] After solving for the dependencies between the parameters of β in the parameter space, the local random parameter β is iteratively optimized using a loss function, which is expressed as:

[0081]

[0082] In the formula, Let y represent the loss function. n Let β represent the nth topological state. -n Let η(y) represent the local random parameter, α represent the global deterministic parameter, and η(y) represent the global deterministic parameter. n ,β -n ,α) represents y n ,β -n The natural parameter of α The parameters representing the variational distribution. Represents η(y) n ,β -n The expected value of ,α).

[0083] 2. Solve for the part related to α.

[0084] Based on empirical Bayesian point estimation, given β, the optimal setting of α can be obtained by maximizing the marginal likelihood p(Y|X,α,β). Considering that p(Y|X,α,β) cannot be directly calculated, a variational expectation-maximization method is used to maximize... To optimize α.

[0085]

[0086] In the formula, ξ represents the step size, and α represents the global deterministic parameter. This represents the lower bound of evidence for optimizing α. This represents the optimized estimate of α.

[0087] After optimizing the local random parameter β and the global deterministic parameter α, stochastic variational inference is introduced. Combined with a mini-batch sampling strategy, a stochastic optimization method with noisy natural gradients is used to optimize the distribution inference objective function. The local random parameter β and the variational distribution parameter φ are initialized using prior knowledge. The following steps are repeated: a mini-batch D is randomly sampled from D. B D represents the input data of the preset topology prediction model; calculate the estimated values ​​of the variational distribution parameter φ and the optimized α. Update the global deterministic parameter α to The weighted average of the current values ​​is calculated, and the weights ω are updated. The iteration continues until the maximum iteration step size is reached, then the optimized parameters α, β, and φ are returned.

[0088] Step 1022: Receive the generator active power, generator reactive power, load active power, load reactive power, topology status, and meteorological data at the current time in the input layer. Perform a linear transformation on the generator active power, generator reactive power, load active power, load reactive power, topology status, and meteorological data at the next time in the hidden layer to obtain sample features. In the output layer, convert the sample features into the topology status of the power system at the next time based on the posterior distribution.

[0089] In this embodiment, a three-layer topology prediction model is constructed to process the current generator active power, current generator reactive power, current load active power, current load reactive power, current power system topology state, and next-time meteorological data to obtain the power system topology state at the next moment. In the input layer, the power system obtains key parameters from the current moment, including the active and reactive power of generators and loads, the power system topology state, and future meteorological data. These data collectively influence the power system topology state. In the hidden layer, linear transformations are applied to process the input data, extracting sample features that reflect the characteristics of the power system. These sample features contain potential information about power system topology state transitions. In the output layer, posterior distribution and Bayesian inference techniques are used to transform the obtained sample features into predictions of the future power system topology state and its probability of occurrence. This provides a data-driven prediction mechanism for the power system, accurately identifying potential changes, mitigating potential risks, and improving the stability and reliability of power system operation.

[0090] For example, such as Figure 2 The diagram shows the network structure of the topology prediction model. The input layer X receives multiple input data x1, x2...x k(i.e., the current generator active power, the current generator reactive power, the current load active power, the current load reactive power, the current power system topology, and the meteorological data for the next moment), the topology prediction model includes local stochastic parameters β and global deterministic parameters α (i.e., α1...α2). m The globally deterministic parameter α plays a role in the input data (i.e., linear transformation) in the input layer and multiple hidden layers, while the locally random parameter β plays a role in the output hidden layer in conjunction with the posterior distribution. The output layer Y outputs the predicted topological state of the power system at the next time step (i.e., y1...y2). n ).

[0091] Step 103: Input the data samples from multiple historical moments, the topology status of the next moment, and the meteorological data of the next moment into the preset operation status prediction model to calculate the bus voltage and load rate of the next moment.

[0092] In this embodiment, by inputting historical data samples, predicted topology states, and future meteorological data into a pre-set operational state prediction model, the bus voltage and load factor of the power system at the next moment can be accurately predicted. Historical data samples include data samples of the power system at the current moment and over a past period, providing the operational state prediction model with temporal information about the power system's operation. The topology state at the next moment reflects the layout and connection methods of the power system's components, directly affecting the flow and distribution of electricity. Meteorological data can influence electricity demand and generation efficiency, especially under extreme weather conditions. Combining these factors allows for a more comprehensive understanding of the dynamic characteristics of the power system, thereby predicting its stability. By predicting bus voltage and load factor, potential risks can be identified in advance, determining whether the power system faces risks of overload, faults, or instability. This provides crucial information for optimized operation and risk management of the power system, improving its safety, reliability, and efficiency, enabling timely early warnings, and preventing power outages and other faults.

[0093] For example, such as Figure 3 The diagram shown illustrates the structure of the operational state prediction model, which incorporates data samples from multiple historical moments (i.e., ... The first temporal feature is extracted from each LSTM module in the first layer (LSTM Layer 1) of the two-layer long short-term memory network. This first temporal feature is then input into each LSTM module in the second layer (LSTM Layer 2) to extract the second temporal feature. The topology state and meteorological data for the next time step are input into the Dense module in the first layer (FC Layer 1) of the two-layer fully connected network to extract the first abstract feature. This first abstract feature is then input into the Dense module in the second layer (FC Layer 2) to extract the second abstract feature. The second temporal feature and the second abstract feature are concatenated to obtain the concatenated feature. This concatenated feature is then processed in a pre-defined operating state prediction model, and the output layer outputs two Dense (fully connected layer) outputs of the power system's bus voltage (V) for the next time step. t+1 ) and the load factor at the next time step (i.e., p) t+1 ), where t+1 represents the next moment.

[0094] For training the operational status prediction model, multi-source heterogeneous spatiotemporal data is obtained by collecting historical data from the energy management system or simulation data from multiple platforms. Then, through data preprocessing, feature selection, and data integration, training samples for the operational status prediction model are obtained. The operational status prediction model can be trained offline or online.

[0095] Step 104: For the next time step, calculate the risk value of the power system based on the topology, bus voltage, and load rate.

[0096] In this embodiment, the risk value of the power system is calculated based on the topology, bus voltage, and load rate of the power system at the next moment to assess its stability and security. This process quantifies the potential risks to the power system by comprehensively analyzing the topology, bus voltage, and load rate at the next moment. The topology describes the connection and layout of various components in the power system (such as generators, transformers, and transmission lines), directly affecting the distribution of power flow and the stability of the system. The bus voltage reflects the voltage distribution in the power system; excessively high or low voltage can lead to equipment damage or unstable power supply. The load rate represents the load situation of various parts of the power system; overload can cause equipment failure or system collapse. By combining the topology, bus voltage, and load rate at the next moment, the calculated risk value can identify potential hazards in the power system, such as power overload, equipment failure, or system instability. The calculation of the risk value not only helps identify potential risks in the current system state but also provides a basis for subsequent alarm operations, thereby achieving early warning and preventing major failures.

[0097] For example, the risk value is expressed as:

[0098] risk(t+1,OS i ) = p t+1 (Topo i )*Sev(V t+1 ,p t+1 );

[0099]

[0100] In the formula, risk(t+1, OS) i ) represents the risk indicator, t represents time, and OS i Topo represents the combination of bus voltage and load factor. i p represents the topological state. t+1 (Topo i V represents the probability of the topological state at the next time step. t+1 p represents the bus voltage at the next moment. t+1 Sev(V) represents the load factor of the transmission lines and transformers at the next moment. t+1 ,p t+1 α represents the severity of the relationship between the bus voltage at the next moment and the load rate of the transmission lines and transformers at the next moment. i Let β represent the i-th globally deterministic parameter. i Let represent the j-th local random parameter, N represent the number of globally deterministic parameters, L represent the number of local random parameters, and Sev(V) represent the number of locally random parameters. t+1 ) represents the severity of the bus voltage at the next moment, Sev(p) t+1 The Ω value represents the severity of the load factor of the transmission lines and transformers at the next moment. t This represents a set of topological states. `if` represents a conditional statement; if (V) ... t+1 )<0.9 means when V t+1 When it is less than 0.9, Sev(V) t+1 ) equals 0.9 and V t+1 The difference between them; when 0.9 ≤ if(V t+1 )≤1.1 means when V t+1 Between 0.9 and 1.1, Sev(V t+1 ) equals 0; when if(V t+1 )>1.1 means when V t+1 Greater than 1.1, Sev(V) t+1 ) equals V t+1 The difference between 1 and 1.1; when if(p t+1 )≤0.9 indicates p t+1 Within 0.9, Sev(p t+1 ) equals 0; when if(pt+1 )>0.9 indicates p t+1 Greater than 0.9, Sev(p t+1 ) equals p t+1 The difference between 0.9 and 0.9.

[0101] Step 105: Perform alarm operations on the power system based on the risk value.

[0102] In this embodiment, the risk value reflects the degree of risk that the power system may fail or become unstable in the future. By identifying and responding to potential failures or instability in the power system and triggering alarm operations based on the calculated risk value, the stability and reliability of the power system can be ensured.

[0103] For example, embodiments of the present invention perform alarm operations on the power system in two ways as follows:

[0104] Method 1: If the risk value is greater than or equal to the preset risk threshold, an alarm operation will be performed on the power system.

[0105] Method 2: Based on the risk values ​​at multiple future moments, the system state of the power system is constructed. Multiple consecutive system states are selected according to a preset time step to obtain the state trajectory. By observing the evolution path of the system state, the dynamic changes that may occur in the power system over a long period of time can be captured, thereby identifying possible sudden events or long-term accumulated problems.

[0106] For multiple state trajectories, if a power outage occurs in a state trajectory, the state trajectory is marked with fault information, and the topology state of the power outage is recorded as the target topology. This marks the possible power outages of transmission lines in the state trajectory and records the topology state at the time of the fault, providing accurate background information for future alarm operations. It also provides an important basis for fault cause analysis, prediction and solution formulation.

[0107] The ratio between the number of fault information and the number of state trajectories is calculated as the fault probability. By using the ratio of the number of fault information to the number of state trajectories, the frequency or probability of faults in the power system can be quantified, thereby providing data support for alarm decision-making. If the fault probability is high, it indicates that the power system faces greater risks in the future, and the triggering conditions for alarm operations will be more stringent.

[0108] Alarm operations are performed on the power system based on the target topology and fault probability.

[0109] To better understand this embodiment, the following will be explained... Figure 4 The diagram shown illustrates the structure of a power system risk alarm and demonstrates the execution process of an embodiment of the present invention.

[0110] In the data preprocessing module, a multi-source database (i.e., data samples are constructed at various time points) is formed through the collection and generation of historical data and simulation data. The topology prediction model receives the training set from this multi-source database, including topology data, power trough (PF) data (i.e., tidal data), and meteorological data. This training set includes the current generator active power, current generator reactive power, current load active power, current load reactive power, current power system topology state, and meteorological data for the next time point. This training set is then input into the topology prediction model for training, outputting the power system topology state for the next time point (e.g., Topo1...Topo). N ) and the probability corresponding to the topological state at the next time step (e.g., Prob1....Prob N );

[0111] In the operational status prediction model, data from multiple databases are received to form the training set of the operational status prediction model, which includes topology data, power voltmeter (PF) data, meteorological data, and forecast data. Feature selection and data integration are performed on the training set data to form the operational status prediction model knowledge base. The operational status prediction model knowledge base data and the topology status of the power system at the next moment output by the topology prediction model are input into the operational status prediction model to obtain the bus voltage and load rate of the power system at the next moment.

[0112] In the risk assessment and early warning model, the probabilities corresponding to the bus voltage, load rate, and topology state at the next time step are input into the severity calculation model. The model outputs the risk value (Sev*Prob) of each state (such as state 1...state N) (i.e., the topology state at the next time step), and performs risk assessment and early warning based on the risk value.

[0113] In this embodiment of the invention, data samples are constructed at various times. Each data sample includes bus voltage, generator active power, generator reactive power, load active power, load reactive power, power system topology, load rate of transmission lines and transformers, and meteorological data. The current generator active power, generator reactive power, load active power, load reactive power, topology, and meteorological data for the next time are input into a preset topology prediction model to calculate the topology for the next time. Data samples from multiple historical times, the topology for the next time, and the meteorological data for the next time are input into a preset operating state prediction model to obtain the bus voltage and load rate for the next time. For the next time, the risk value of the power system is calculated based on the topology, bus voltage, and load rate. Based on the risk value, an alarm operation is performed on the power system. By predicting the topology, operating state, and risk value, potential faults can be detected and warned in advance, thereby reducing the risk of power outages and accidents. Meanwhile, the integration of meteorological data enhances the power system's ability to respond to emergencies. By comprehensively considering multiple data sources and sequentially using topology prediction models and operational status prediction models to predict power system risks, the correlation between various data sources is strengthened, and the accuracy of power system risk prediction is improved.

[0114] Example 2

[0115] Figure 5 This is a schematic diagram of the structure of a power system risk alarm device provided in Embodiment 2 of the present invention, as shown below. Figure 5 As shown, the device includes:

[0116] The data sample construction module 501 is used to construct data samples at various times; each of the data samples includes bus voltage, generator active power, generator reactive power, load active power, load reactive power, power system topology, transmission line and transformer load rate, and meteorological data.

[0117] The topology prediction module 502 is used to input the generator active power, generator reactive power, load active power, load reactive power, topology state, and meteorological data at the next moment into a preset topology prediction model to calculate the topology state at the next moment.

[0118] The motion state prediction module 503 is used to input the data samples from multiple historical moments, the topology state at the next moment, and the meteorological data at the next moment into a preset operation state prediction model to calculate the bus voltage and the load rate at the next moment.

[0119] The risk value calculation module 504 is used to calculate the risk value of the power system for the next time step based on the topology state, the bus voltage and the load rate.

[0120] The power system alarm module 505 is used to perform alarm operations on the power system based on the risk value.

[0121] In one embodiment of the present invention, the topology prediction module 502 includes:

[0122] The posterior distribution setting module is used to set the posterior distribution for a preset topology prediction model; the topology model includes an input layer, a hidden layer, and an output layer.

[0123] The input data processing module is used to receive the generator active power, generator reactive power, load active power, load reactive power, topology state, and meteorological data at the current time in the input layer, and at the next time in the hidden layer. It performs a linear transformation on the generator active power, generator reactive power, load active power, load reactive power, topology state, and meteorological data at the next time in the hidden layer to obtain sample features. Finally, it converts the sample features into the topology state of the power system at the next time in the output layer based on the posterior distribution.

[0124] In one embodiment of the present invention, the posterior distribution setting module includes:

[0125] The posterior distribution function construction module is used to construct the posterior distribution function for a pre-defined topological prediction model based on Bayes' theorem.

[0126] The approximate posterior acquisition module is used to perform distribution inference based on the posterior distribution function to obtain an approximate posterior as the posterior distribution of the topological model;

[0127] The posterior distribution function is expressed as:

[0128]

[0129] In the formula, p(θ|D) represents the posterior distribution function, θ represents the network parameters of the topology prediction model, D represents the data sample, p(θ) represents the prior distribution, p(D|θ) represents the likelihood function, and p(D) represents the marginal likelihood. This represents the i-th data sample. This represents the i-th topological state. Indicates in The probability of the occurrence of the i-th topological state under the conditions of θ, where N represents the number of data samples.

[0130] In one embodiment of the present invention, the approximate posterior acquisition module includes:

[0131] The variational distribution setting module is used to set the variational distribution during the distribution inference process of the posterior distribution function;

[0132] The evidence lower bound acquisition module is used to calculate the divergence between the variational distribution and the posterior distribution function to obtain the evidence lower bound.

[0133] The evidence lower bound optimization module is used to optimize the evidence lower bound based on the loss function and Bayesian point estimation to obtain an approximate posterior as the posterior distribution of the topological model.

[0134] The lower bound of the evidence is represented as follows:

[0135]

[0136] In the formula, θ represents the divergence between the variational distribution and the posterior distribution function, and θ represents the network parameters of the topology prediction model. Let p(θ|D) represent the variational distribution, p(D,θ) represent the posterior distribution function, p(D,θ) represent the joint distribution, and p(D) represent the marginal likelihood. express Expected value, IE q [logp(D,θ)] represents the expected value of p(D,θ). Indicates the lower bound of the evidence; const indicates a constant.

[0137] The loss function is expressed as:

[0138]

[0139] In the formula, Let y represent the loss function. n β represents the nth topological state. -n Let η(y) represent the local random parameter, α represent the global deterministic parameter, and η(y) represent the global deterministic parameter. n ,β -n ,α) represents y n ,β -n The natural parameter of α The parameters representing the variational distribution are... Represents η(y) n ,β -n The expected value of ,α).

[0140] In one embodiment of the present invention, the running state prediction model includes a two-layer long short-term memory network and a two-layer fully connected network; the motion state prediction module 503 includes:

[0141] The first temporal feature acquisition module is used to input the data samples from multiple historical moments into the first layer of the two-layer long short-term memory network in chronological order to extract the first temporal features;

[0142] The second temporal feature acquisition module is used to input the first temporal feature into the second layer of the two-layer long short-term memory network and extract the second temporal feature.

[0143] The first abstract feature acquisition module is used to input the topological state at the next time step and the meteorological data at the next time step into the first layer of the two-layer fully connected network to extract the first abstract feature.

[0144] The second abstract feature acquisition module is used to input the first abstract feature into the second layer of the two-layer fully connected network and extract the second abstract feature.

[0145] The feature concatenation module is used to concatenate the second temporal feature with the second abstract feature to obtain the concatenated feature;

[0146] The splicing feature calculation module is used to calculate the splicing features in a preset operating state prediction model to obtain the bus voltage and the load rate at the next time step.

[0147] In one embodiment of the present invention, the risk value is represented as:

[0148] risk(t+1,OS i ) = p t+1 (Topo i )*Sev(V t+1 ,p t+1 );

[0149]

[0150] In the formula, risk(t+1, OS) i ) represents the risk indicator, t represents time, and OS i This represents the combination of the bus voltage and the load rate, Topo i p represents the topological state. t+1 (Topo i V represents the probability of the topological state at the next moment. t+1 p represents the bus voltage at the next moment. t+1 Sev(V) represents the load rate of the transmission line and transformer at the next moment. t+1 ,pt+1 α represents the severity of the relationship between the bus voltage at the next time step and the load rate of the transmission line and transformer at the next time step. i Let β represent the i-th globally deterministic parameter. i Let represent the j-th local random parameter, N represent the number of global deterministic parameters, L represent the number of local random parameters, and Sev(V) represent the number of locally random parameters. t+1 ) represents the severity of the bus voltage at the next moment, Sev(p) t+1 The value Ω represents the severity of the load rate of the transmission line and transformer at the next moment. t The set represents the topological states, and if represents a conditional statement.

[0151] In one embodiment of the present invention, the power system alarm module 505 includes:

[0152] The risk value judgment module is used to perform an alarm operation on the power system if the risk value is greater than or equal to a preset risk threshold.

[0153] And / or,

[0154] The system state construction module is used to construct the system state of the power system based on the risk values ​​at multiple future time points;

[0155] The state trajectory acquisition module is used to filter out multiple consecutive system states according to a preset time step to obtain a state trajectory.

[0156] The target topology recording module is used to mark fault information on the state trajectories if a power outage occurs in one of the state trajectories, and to record the topology state in which the power outage occurs as the target topology.

[0157] The fault probability calculation module is used to calculate the ratio between the number of fault information and the number of state trajectories as the fault probability.

[0158] The alarm operation module is used to perform alarm operations on the power system based on the target topology and the fault probability.

[0159] The power system risk alarm device provided in this embodiment of the invention can execute the power system risk alarm method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of executing the power system risk alarm method.

[0160] Example 3

[0161] See Figure 6This diagram illustrates a structural schematic of a computer device according to an embodiment of the present invention. The term "computer device" is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, blade servers, mainframe computers, and other suitable computers. The computer device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0162] like Figure 6 As shown, the computer device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer programs stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the computer device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0163] Multiple components in computer device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of monitors, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows computer device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0164] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as power system risk alarm methods.

[0165] In some embodiments, the power system risk alarm method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on computer device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the power system risk alarm method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the power system risk alarm method by any other suitable means (e.g., by means of firmware).

[0166] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0167] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0168] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0169] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer device having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0170] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0171] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0172] Example 4

[0173] This invention also provides a computer program product, which includes a computer program that, when executed by a processor, implements the power system risk alarm method provided in any embodiment of this invention.

[0174] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0175] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0176] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A power system risk alarm method, characterized in that, include: Data samples are constructed at various times; each data sample includes bus voltage, generator active power, generator reactive power, load active power, load reactive power, power system topology, load rate of transmission lines and transformers, and meteorological data. The current generator active power, the current generator reactive power, the current load active power, the current load reactive power, the current topology state, and the meteorological data for the next time are input into a preset topology prediction model to calculate the topology state for the next time. The data samples from multiple historical moments, the topology state at the next moment, and the meteorological data at the next moment are input into a preset operation state prediction model to calculate the bus voltage and the load rate at the next moment. For the next moment, calculate the risk value of the power system based on the topology state, the bus voltage, and the load rate; An alarm operation is performed on the power system based on the risk value; The step of inputting the generator active power, generator reactive power, load active power, load reactive power, topology state, and meteorological data for the next moment into a preset topology prediction model to calculate the topology state for the next moment includes: A posterior distribution is set for a pre-defined topology prediction model; the topology prediction model includes an input layer, a hidden layer, and an output layer. The input layer receives the current generator active power, the current generator reactive power, the current load active power, the current load reactive power, the current topology state, and the meteorological data for the next time. The hidden layer performs a linear transformation on the current generator active power, the current generator reactive power, the current load active power, the current load reactive power, the current topology state, and the meteorological data for the next time to obtain sample features. The output layer converts the sample features into the topology state of the power system for the next time according to the posterior distribution. The operational status prediction model includes a two-layer long short-term memory network and a two-layer fully connected network; the step of inputting data samples from multiple historical moments, the topology state at the next moment, and the meteorological data at the next moment into the preset operational status prediction model for calculation to obtain the bus voltage and the load rate at the next moment includes: The data samples from multiple historical moments are input into the first layer of the two-layer long short-term memory network in chronological order to extract the first temporal features; The first temporal feature is input into the second layer of the two-layer long short-term memory network to extract the second temporal feature; The topological state at the next time step and the meteorological data at the next time step are input into the first layer of the two-layer fully connected network to extract the first abstract feature; The first abstract feature is input into the second layer of the two-layer fully connected network to extract the second abstract feature; The second temporal feature is concatenated with the second abstract feature to obtain the concatenated feature; The splicing features are calculated in a preset operating state prediction model to obtain the bus voltage and the load rate at the next time step.

2. The method according to claim 1, characterized in that, Setting the posterior distribution for the preset topology prediction model includes: Based on Bayes' theorem, a posterior distribution function is constructed for the pre-defined topological prediction model; Based on the posterior distribution function, distribution inference is performed to obtain an approximate posterior of the posterior distribution of the topology prediction model; The posterior distribution function is expressed as: In the formula, p(θ|D) represents the posterior distribution function, θ represents the network parameters of the topology prediction model, D represents the data sample, p(θ) represents the prior distribution, p(D|θ) represents the likelihood function, and p(D) represents the marginal likelihood. This represents the i-th data sample. This represents the i-th topological state. Indicates in The probability of the occurrence of the i-th topological state under the conditions of θ, where N represents the number of data samples.

3. The method according to claim 2, characterized in that, The step of performing distribution inference based on the posterior distribution function to obtain an approximate posterior distribution as the posterior distribution of the topology prediction model includes: In the process of performing distribution inference on the posterior distribution function, a variational distribution is set; Calculate the divergence between the variational distribution and the posterior distribution function to obtain the lower bound of evidence; The lower bound of evidence is optimized based on the loss function and Bayesian point estimation to obtain an approximate posterior as the posterior distribution of the topological prediction model; The lower bound of the evidence is represented as follows: In the formula, θ represents the divergence between the variational distribution and the posterior distribution function, and θ represents the network parameters of the topology prediction model. Let p(θ|D) represent the variational distribution, p(D,θ) represent the posterior distribution function, p(D,θ) represent the joint distribution, and p(D) represent the marginal likelihood. express Expected value, IE q [logp(D,θ)] represents the expected value of p(D,θ). Indicates the lower bound of the evidence; const indicates a constant. The loss function is expressed as: In the formula, Let y represent the loss function. n β represents the nth topological state. -n Let η(y) represent the local random parameter, α represent the global deterministic parameter, and η(y) represent the global deterministic parameter. n ,β -n ,α) represents y n ,β -n The natural parameter of α The parameters representing the variational distribution are... Represents η(y) n ,β -n The expected value of ,α).

4. The method according to any one of claims 1-3, characterized in that, The risk value is expressed as: risk(t+1,OS i )=p t+1 (Topo i )*North(V t+1 ,p t+1 ); In the formula, risk(t+1, OS) i ) represents the risk indicator, t represents time, and OS i This represents the combination of the bus voltage and the load rate, Topo i p represents the topological state. t+1 (Topo I V represents the probability of the topological state at the next moment. t+1 p represents the bus voltage at the next moment. t+1 Sev(V) represents the load rate of the transmission line and transformer at the next moment. t+1 ,p t+1 α represents the severity of the relationship between the bus voltage at the next time step and the load rate of the transmission line and transformer at the next time step. i Let β represent the i-th globally deterministic parameter. i Let represent the j-th local random parameter, N represent the number of global deterministic parameters, L represent the number of local random parameters, and Sev(V) represent the number of locally random parameters. t+1 ) represents the severity of the bus voltage at the next moment, Sev(p) t+1 The value Ω represents the severity of the load rate of the transmission line and transformer at the next moment. t The set represents the topological states, and if represents a conditional statement.

5. The method according to claim 4, characterized in that, The method of issuing an alarm to the power system based on the risk value includes: If the risk value is greater than or equal to the preset risk threshold, an alarm operation is performed on the power system. And / or, The power system is constructed based on the risk values ​​at multiple future points in time; Multiple consecutive system states are selected according to a preset time step to obtain a state trajectory; For multiple state trajectories, if a power outage occurs in a power transmission line in a state trajectory, the state trajectory is marked with fault information, and the topology state in which the power outage occurs is recorded as the target topology. The ratio between the number of fault information entries and the number of state trajectories is calculated as the fault probability. The power system is alarmed based on the target topology and the fault probability.

6. A computer device, characterized in that, The computer device includes: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the power system risk alarm method as described in any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the power system risk alarm method as described in any one of claims 1-5.

8. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the power system risk alarm method as described in any one of claims 1-5.

Images