Markov model modeling method for automatically carrying out intelligent power grid monitoring

Abstract

The invention belongs to the technical field of power distribution automation, and particularly relates to a Markov model modeling method for automatically performing intelligent power grid monitoring, which analyzes power grid topology, fault occurrence, load change and power failure conditions through a state transition probability, realizes accurate prediction and state optimization, and adopts exploration initialization and utilization optimization strategies to realize intelligent power grid monitoring. And in combination with a self-adaptive exploration-utilization balance adjustment strategy, the system can dynamically optimize a decision in real time according to the power grid. The method has remarkable advantages in the aspects of load management and energy scheduling optimization, energy distribution can be optimized through load fluctuation analysis, the deployment efficiency of the energy storage units is improved, energy loss is reduced, demand response is optimized, and therefore the economical efficiency of a power system and the energy supply stability are improved. A Markov model is used for network security monitoring, potential network attacks, electric power fraud and data tampering are detected, the power grid defense capability is improved, and a security policy is optimized in combination with a vulnerability scanning technology.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid monitoring, and in particular relates to a Markov modeling method for automatically performing smart grid monitoring. Background Art

[0002] With the growing penetration of smart grids, distribution automation (DA) plays a key role in addressing these challenges by providing advanced monitoring and control systems. Transition probabilities between states, driven by factors such as grid topology, fault occurrence, load variations, and outages, can be captured using Markov models. This helps predict and analyze grid behavior under varying operating conditions. However, existing Markov models are limited to specific aspects of smart grids, such as load forecasting, fault and intrusion detection, and there is no unified Markov modeling approach that encompasses all of these aspects. Second, Markov models have not been tested with real-world data. Using simulated data can lead to overfitting and inaccurate results, as these data values do not represent the real world. Third, currently proposed models have shortcomings in handling noise and uncertainty. Smart grid data often exhibits noise and uncertainty, which, if not accounted for, can degrade model performance. Finally, currently developed Markov models lack scalability. As smart grids continue to expand, become increasingly complex, and become more interconnected, existing approaches may struggle to cope with these changes. To address these research gaps, it is imperative to develop a unified Markov model framework for smart grid monitoring.

[0003] Existing smart grid monitoring technologies suffer from numerous shortcomings in fault detection, load management, security, and market optimization. First, traditional methods rely on fixed rules or statistical analysis for fault detection, making it difficult to accurately predict faults, resulting in low detection efficiency and long recovery times. Second, load management and energy scheduling optimization capabilities are limited, making them unable to effectively address fluctuating load demand, leading to low energy utilization and high scheduling costs. Furthermore, existing grid security mechanisms lack sufficient defenses against cyberattacks, power fraud, and data tampering, resulting in low security. Market pricing strategies still rely on static models, making it difficult to adapt to the dynamic changes in the energy market, resulting in low transaction efficiency and insufficient cost optimization. Overall, current technologies lack intelligence, adaptability, and optimization efficiency, making them unable to meet the demands of modern power grids for efficient, stable, and secure operation. Summary of the Invention

[0004] In order to solve the above problems, the present invention provides a Markov modeling method for automated smart grid monitoring. The specific technical solution is as follows:

[0005] A Markov modeling method for automated smart grid monitoring includes the following steps: S1, describing the probability transition relationship of states over time in distributed automation-based smart grid monitoring;

[0006] S2. Comprehensive metrics for evaluating Markov models for smart grid monitoring and distributed automation;

[0007] S3, represents the state transition probability of the distribution automation system;

[0008] S4, describing the probability transfer relationship of smart grid load changes;

[0009] S5. Describe the state φ of the distribution automation system from time t t Transition to state φ at time t+1 t+1 probability;

[0010] S6. Determine the daily load demand percentage;

[0011] S7, describing the dynamic change of the state of charge of the kth energy storage unit in the smart grid;

[0012] S8. Establish a fault detection and service restoration model in smart grid monitoring;

[0013] S9, calculating the power quality index at time t;

[0014] S10. Evaluate the impact of power outages and customer satisfaction;

[0015] S11. Describe the comprehensive assessment of smart grid security using Markov models.

[0016] S12. Evaluate smart grid monitoring system based on Markov model and distribution automation.

[0017] Preferably, the step S1 is implemented by the following mathematical model: The future state depends only on the current state, not on the previous state sequence. In this model, t represents the state of the smart grid system at time t, where ξ t Represents a specific state in the state set S. This equation means that if the system is currently in state ξ t , then it transfers to state ξ at time t+1 t+1 The probability of the transition is the probability of all possible previous states ξ t-1 ∈S and sum the results, and press t-1 Transfer to ξ t , and then from ξ t Transfer to ξ t+1 The joint probability of is weighted;

[0018] In the step S2, it is achieved by the following mathematical model: Where Φ represents an aggregated metric that combines the impact of energy reliability and financial consequences over a specific time period T. The integral is from 0 to T, capturing the cumulative effect of dynamic factors that affect energy supply reliability and related financial costs. In this equation, λ(t) and μ(t) are weighting functions used to adjust the unsupplied energy (E NS (t)) and financial impact (C F (t)) importance, E total (t) represents the total energy demand or capacity at time t, and provides a reference for evaluating the proportion of unsupplied energy to total demand. max Acts as a normalizing factor to ensure that the financial impact is assessed relative to the maximum allowed threshold.

[0019] Preferably, step S3 is implemented by the following mathematical model:

[0020] Indicates that the system changes from the state φ at time t t Transition to state φ at time t+1 t+1 The possibility of configuring s in a given state t Under the condition of t Indicates the current state of the distribution automation system, which is affected by the variable D t Influence, variable S t Contains properties related to the system operating settings and environmental conditions. This equation is valid for all possible configurations D t Sum up, each configuration represents a different operating scenario that affects state transition, conditional probability P(Φ t+1 =φ t+1 ∣Φ t =φ t ,S t =s t ,D t =dt) quantifies the t In the case of transfer to φ t+1 probability;

[0021] In the step S4, it is achieved by the following mathematical model:

[0022] At time t, it is in state λ t Consider the load characteristics After that, the system transfers to state λ at time t+1 t+1 The probability of t Indicates the current load change state in the smart grid system, which is affected by L t Variable influence, L tIncluding peak load, average load, minimum load, load curve and load composition parameters, the equation sums all possible load profiles P. Each load profile represents a different operating configuration and affects the state transition. The conditional probability P(Λ t+1 =λ t+1 ∣Λ t =λ t , P t =l t ) quantifies the system shift to λ t+1 possibility;

[0023] In the step S5, it is achieved by the following mathematical model:

[0024] Consider the current state φ t and fault characteristics f t , which sums all possible failure scenarios R, which affect the probability P(Φ t+1 =φt+1|Φ t =φ t ,F t =f t ,R t =f t ) This equation quantifies the impact of different fault rates, types, and locations on system behavior and is crucial for optimizing fault management strategies in smart grids.

[0025] Preferably, step S6 is implemented by the following mathematical model: It is crucial for the effectiveness of the proposed Markov model in smart grid management, where P(t) represents the percentage of total load demand at time t, which is calculated by subtracting the energy demand E from all N Markov models. i The sum of (t) and the distribution automation factor D in all M models j (t) and multiplied by 100. This formula is the cornerstone of using Markov models to evaluate and optimize energy distribution strategies. These models are an important part of the field of smart grid monitoring. With their inherent ability to explore and exploit system dynamics, they promote adaptive and efficient operation modes. In practical applications, each E i (t) represents the energy demand of the i-th Markov model at a certain time, reflecting the dynamic fluctuation of energy usage pattern, while D j (t) represents the distribution automation factor of the jth Markov model, reflecting the degree of automation control and optimization in the energy distribution process. By summarizing the demands and factors in multiple models, this equation provides a global view of the overall load demand over time;

[0026] In the step S7, it is achieved by the following mathematical model: Among them, S k (t) represents the SoC state at time t, which is updated based on the current state, the energy output E from the N Markov models i (t) and the distribution automation factor D from M models j (t), parameter α k , β k and γ k The equation determines the charging and discharging dynamics and efficiency of the energy storage unit. It illustrates how the energy storage unit can be combined with the Markov model to optimize grid stability and energy distribution efficiency using exploration-exploitation strategies.

[0027] Preferably, in step S8, the following mathematical model is used: ij (t) = α ij ·V ij (t)+β ij ·C ij (t)-γ ij ·F ij-1 (t)+δ ij ·D ij (t), where F ij (t) represents the probability of event i and Markov model j failing at time t, affected by voltage sag / swell (V ij (t)), current spike (C ij (t)), previous fault status (F ij-1 (t)) and the distortion percentage (D ij (t)). Parameter α ij , β ij , γ ij and δ ij The impact of these factors on fault probability and service restoration decisions is determined. The equation illustrates how Markov models can integrate multiple metrics to enhance distribution automation capabilities and maintain grid stability and reliability under different electrical disturbance conditions.

[0028] In the step S9, it is achieved by the following mathematical model: Its measurement range covers 0 to 360 degrees simultaneously. The index combines the energy demand E i (t) In the distribution automation factor D j The weighted average value above (t), and the number of failures F n (t) In the impact factor G pThe index provides a comprehensive measure of grid stability and efficiency based on the trade-off calculation above (t). Influenced by the exploration-exploitation strategies in the Markov model, it promotes real-time evaluation and optimization in smart grid monitoring and automation, which is crucial for maintaining reliable power services and minimizing interference.

[0029] Preferably, the step S10 is implemented by the following mathematical model: Through the energy demand E i (t) In the power distribution factor D j The weighted calculation above (t), the number of failures F k (t) in factor G l (t) above, and the operating factor H m (t)In reliability index I n Calculations above (t).

[0030] In the step S11, the following mathematical model is used: Among them, α i , β i and γ i represents the parameters that control different stages, the parameter δ represents the key exploration-exploitation trade-off factor, and j It represents the hybrid exploration-exploitation strategy. This formula calculates the index Φ Grid , used to evaluate the effectiveness of vulnerability detection mechanisms;

[0031] In the step S12, it is achieved by the following mathematical model: Among them, E represents the overall effectiveness of the entire system in the time period T, the parameters α and (1-α) represent the weights given to the exploration strategy and the utilization strategy respectively, and P explore (t) and P exploit (t) represents the corresponding probability or score at time t, the ratio represents a cost-effectiveness measure, where C(t) and P(t) are the energy production cost and consumer price at time t, respectively, and the exponent β regulates the system's sensitivity to this cost-price ratio.

[0032] Preferably, in step S12, the Markov model optimization method is as follows:

[0033] S121, enter from the start module, and set the initial parameters of simulated annealing and reinforcement learning in the initialization module;

[0034] S122, enter the exploration phase. In the exploration initialization module, simulated annealing is allowed to widely explore the solution space, generate diverse initial solutions, avoid local optimal solutions, and discover potential global optimal solutions. Pre-training is used in the exploration-exploitation trade-off parameter module to determine the acceptance probability and exploration intensity of simulated annealing and the balance between exploring new states and exploiting existing knowledge in reinforcement learning.

[0035] S123. Utilize the optimization module to refine the solutions found in the exploration phase using reinforcement learning and focus on promising areas in the search space to improve the quality of the solution and the speed of convergence.

[0036] S124, using an adaptive exploration-exploitation balance module to monitor the convergence rate and performance indicators and dynamically adjust the balance between exploration and exploitation;

[0037] S125, measuring the performance of the optimized solution in the solution quality evaluation module and evaluating the reliability and efficiency of smart grid monitoring;

[0038] S126. In the convergence check module, ensure whether the optimization process meets the convergence criteria. If the convergence criteria are met, enter the optimal solution finding module, the algorithm outputs the optimized solution and ends the optimization process. If the convergence criteria are not met, the algorithm will check whether the iteration upper limit module is reached. If the iteration upper limit has been reached, the optimization process is terminated; if the iteration upper limit has not been reached, the algorithm returns to the exploration stage and continues the optimization.

[0039] S127. Throughout the entire process, the algorithm dynamically adjusts the exploration and development work in the module based on the progress and convergence indicators. At this time, the parameter update module adaptively adjusts the simulated annealing scheduling and reinforcement learning parameters to further optimize the exploration-exploitation process.

[0040] S128. Then, in the continuous optimization module of exploration-exploitation trade-off, the exploration-exploitation trade-off parameters are further refined to ensure the balance of the method during the optimization process.

[0041] S129. A hybrid simulated annealing-reinforcement learning approach is used in the hybrid strategy execution module: simulated annealing is responsible for exploration and generating diverse solutions; reinforcement learning is responsible for utilization and optimization of the solutions and convergence to the optimal solution. The optimization process is completed in the end module.

[0042] The beneficial effects of the present invention are:

[0043] A Markov model was used to optimize smart grid monitoring, achieving significant technical results in several areas. First, the model was able to optimize energy supply scheduling and reduce losses due to energy non-supply (ENS). For example, the exploration initialization model provided subsidies when ENS was low, but as ENS increased, this gradually shifted to economic losses, while other optimization models were better able to adapt to ENS changes, thereby reducing losses. Furthermore, the model demonstrated strong adaptability in terms of grid topology and load management. Different models were suitable for grids of different sizes; the exploration initialization model was suitable for small-scale grids, while the adaptive exploration-exploitation model could be used for larger-scale grids. At the same time, the load change management capability was improved, enabling the grid to stably cope with fluctuations from 15kW to 1000kW.

[0044] In terms of grid fault detection and service restoration, the Markov model improves fault detection speed and restoration efficiency, including the rapid identification and repair of voltage drops, current spikes, frequency fluctuations, and grid distortion, thereby enhancing grid stability. The model also optimizes outage management and improves grid reliability. In particular, the hybrid exploration-exploitation model maintained 100% customer satisfaction even after 80 outages, demonstrating the stability and adaptability of smart grids in outage management. Furthermore, in terms of grid security, the model excels in vulnerability detection. The hybrid exploration-exploitation model achieved a detection rate of up to 90% at security scan level 1, effectively reducing the security risks facing the grid.

[0045] In terms of energy production and cost management, the model optimizes the balance between energy production costs and user electricity prices. The hybrid exploration-exploitation model maintains lower consumer electricity prices when production costs are high, enabling grid companies to provide more economical electricity while ensuring profitability. Furthermore, compared with existing methods, the Markov model demonstrated advantages across several key metrics, including a 7.5% reduction in energy consumption, a 15.5% increase in renewable energy penetration, a 62.1% reduction in outage recovery time, a 32.6% reduction in successful cyberattacks, and an 8.9% improvement in market efficiency. These results demonstrate that the Markov model not only improves grid operational efficiency but also enhances security and market adaptability, making it a valuable tool for smart grid optimization. The model can be combined with artificial intelligence and machine learning technologies to further enhance predictive capabilities and intelligence, promoting the automated management and sustainable development of smart grids.

[0046] This invention utilizes a Markov model-based smart grid monitoring and distribution automation technology. By analyzing grid topology, fault occurrence, load changes, and power outages through state transition probabilities, it achieves accurate prediction and state optimization. It employs an exploration initialization and utilization optimization strategy to improve grid monitoring accuracy and reliability. Combined with an adaptive exploration-utilization balancing adjustment strategy, the system optimizes decisions in real time based on grid dynamics, improving fault response speed and stability. This method offers significant advantages in load management and energy scheduling optimization. It can optimize energy distribution through load fluctuation analysis, improve the deployment efficiency of energy storage units, reduce energy losses, and optimize demand response, thereby enhancing the economic efficiency and energy supply stability of the power system. Furthermore, this technology can optimize market pricing and energy trading management in smart grids. By analyzing the relationship between energy production costs and consumer electricity prices, it maximizes cost-effectiveness, optimizes market integration strategies, and improves energy trading efficiency. Furthermore, the patent addresses enhancing smart grid security, utilizing Markov models for network security monitoring to detect potential cyberattacks, power fraud, and data tampering, thereby enhancing grid defense capabilities. Furthermore, vulnerability scanning technology is incorporated to optimize security strategies. The patent also covers grid health management and asset monitoring, extending equipment life and reducing O&M costs through optimized maintenance plans. It also reduces outage duration and improves customer satisfaction by optimizing service restoration times. In terms of fault detection and recovery, the Markov model can monitor voltage sags, current surges, frequency fluctuations, and grid distortion, improving fault detection accuracy and accelerating recovery, ensuring stable grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly describes the drawings required for the specific embodiments or the description of the prior art. Similar elements or parts are generally identified by similar reference numerals throughout the drawings. Elements or parts in the drawings are not necessarily drawn to scale.

[0048] Figure 1 Flowchart of Markov model in smart grid monitoring;

[0049] Figure 2 Diagram for quality incentive analysis of the proposed Markov model in smart grid monitoring;

[0050] Figure 3 Application of Markov models in smart grid monitoring for fault detection and service restoration.

[0051] Among them, (a) voltage sag / swell (b) current mutation (c) frequency fluctuation (d) distortion rate;

[0052] Figure 4Evaluation of the proposed Markov model for smart grid monitoring using distribution automation Health Management Histogram;

[0053] Figure 5 Grid security diagram for the proposed Markov model for smart grid monitoring using distribution automation;

[0054] Figure 6 Market integration of the proposed Markov model for smart grid monitoring using distribution automation. DETAILED DESCRIPTION

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0056] A specific embodiment of the present invention provides a Markov modeling method for automated smart grid monitoring, comprising the following steps: S1, describing the probability transition relationship of states over time in distributed automation-based smart grid monitoring;

[0057] S2. Comprehensive metrics for evaluating Markov models for smart grid monitoring and distributed automation;

[0058] S3, represents the state transition probability of the distribution automation system;

[0059] S4, describing the probability transfer relationship of smart grid load changes;

[0060] S5. Describe the state φ of the distribution automation system from time t t Transition to state φ at time t+1 t+1 probability;

[0061] S6. Determine the daily load demand percentage;

[0062] S7, describing the dynamic change of the state of charge of the kth energy storage unit in the smart grid;

[0063] S8. Establish a fault detection and service restoration model in smart grid monitoring;

[0064] S9, calculating the power quality index at time t;

[0065] S10. Evaluate the impact of power outages and customer satisfaction;

[0066] S11. Describe the comprehensive assessment of smart grid security using Markov models.

[0067] S12, evaluate the smart grid monitoring system based on Markov model and distribution automation. In step S1, this is achieved through the following mathematical model: The future state depends only on the current state, not on the previous state sequence. In this model, t represents the state of the smart grid system at time t, where ξ t Represents a specific state in the state set S. This equation means that if the system is currently in state ξ t , then it transfers to state ξ at time t+1 t+1 The probability of the transition is the probability of all possible previous states ξ t-1 ∈S and sum the results, and press t-1 Transfer to ξ t , and then from ξ t Transfer to ξ t+1 The joint probability of is weighted;

[0068] In step S2, it is achieved through the following mathematical model: Where Φ represents an aggregated metric that combines the impact of energy reliability and financial consequences over a specific time period T. The integral is from 0 to T, capturing the cumulative effect of dynamic factors that affect energy supply reliability and related financial costs. In this equation, λ(t) and μ(t) are weighting functions used to adjust the unsupplied energy (E NS (t)) and financial impact (C F (t)) importance, E total (t) represents the total energy demand or capacity at time t, and provides a reference for evaluating the proportion of unsupplied energy to total demand. max Acts as a normalizing factor to ensure that the financial impact is assessed relative to the maximum allowed threshold.

[0069] In step S3, it is achieved through the following mathematical model:

[0070] Indicates that the system changes from the state φ at time t t Transition to state φ at time t+1 t+1 The possibility of configuring s in a given state t Under the condition of t Indicates the current state of the distribution automation system, which is affected by the variable D t Influence, variable S t Contains properties related to the system operating settings and environmental conditions. This equation is valid for all possible configurations D t Sum up, each configuration represents a different operating scenario that affects state transition, conditional probability P(Φ t+1 =φ t+1∣Φ t =φ t ,S t =s t ,D t =dt) quantifies the t In the case of transfer to φ t+1 probability;

[0071] In step S4, it is achieved through the following mathematical model:

[0072] At time t, it is in state λ t Consider the load characteristics After that, the system transfers to state λ at time t+1 t+1 The probability of t Indicates the current load change state in the smart grid system, which is affected by L t Variable influence, L t Including peak load, average load, minimum load, load curve and load composition parameters, the equation sums all possible load profiles P. Each load profile represents a different operating configuration and affects the state transition. The conditional probability P(Λ t+1 =λ t+1 ∣Λ t =λ t , P t =l t ) quantifies the system shift to λ t+1 possibility;

[0073] In step S5, it is achieved through the following mathematical model:

[0074] Consider the current state φ t and fault characteristics f t , which sums all possible failure scenarios R, which affect the probability P(Φ t+1 =φt+1|Φ t =φ t ,F t =f t ,R t =f t ) This equation quantifies the impact of different fault rates, types, and locations on system behavior and is crucial for optimizing fault management strategies in smart grids.

[0075] In step S6, it is achieved through the following mathematical model: It is crucial for the effectiveness of the proposed Markov model in smart grid management, where P(t) represents the percentage of total load demand at time t, which is calculated by subtracting the energy demand E from all N Markov models. i The sum of (t) and the distribution automation factor D in all M models j (t) and multiplied by 100. This formula is the cornerstone of using Markov models to evaluate and optimize energy distribution strategies. These models are an important part of the field of smart grid monitoring. With their inherent ability to explore and exploit system dynamics, they promote adaptive and efficient operation modes. In practical applications, each E i (t) represents the energy demand of the i-th Markov model at a certain time, reflecting the dynamic fluctuation of energy usage pattern, while D j (t) represents the distribution automation factor of the jth Markov model, reflecting the degree of automation control and optimization in the energy distribution process. By summarizing the demands and factors in multiple models, this equation provides a global view of the overall load demand over time;

[0076] In step S7, it is achieved through the following mathematical model: Among them, S k (t) represents the SoC state at time t, which is updated based on the current state, the energy output E from the N Markov models i (t) and the distribution automation factor D from M models j (t), parameter α k , β k and γ k The equation determines the charging and discharging dynamics and efficiency of the energy storage unit. It illustrates how the energy storage unit can be combined with the Markov model to optimize grid stability and energy distribution efficiency using exploration-exploitation strategies.

[0077] In step S8, the following mathematical model is used: ij (t) = α ij ·V ij (t)+β ij ·C ij (t)-γ ij ·F ij-1 (t)+δ ij ·D ij (t), where F ij (t) represents the probability of event i and Markov model j failing at time t, affected by voltage sag / swell (V ij (t)), current spike (C ij (t)), previous fault status (Fij-1 (t)) and the distortion percentage (D ij (t)). Parameter α ij , β ij , γ ij and δ ij The impact of these factors on fault probability and service restoration decisions is determined. The equation illustrates how Markov models can integrate multiple metrics to enhance distribution automation capabilities and maintain grid stability and reliability under different electrical disturbance conditions.

[0078] In step S9, it is achieved through the following mathematical model: Its measurement range covers 0 to 360 degrees simultaneously. The index combines the energy demand E i (t) In the distribution automation factor D j The weighted average value above (t), and the number of failures F n (t) In the impact factor G p The index provides a comprehensive measure of grid stability and efficiency based on the trade-off calculation above (t). Influenced by the exploration-exploitation strategies in the Markov model, it promotes real-time evaluation and optimization in smart grid monitoring and automation, which is crucial for maintaining reliable power services and minimizing interference.

[0079] In step S10, the following mathematical model is used: Through the energy demand E i (t) In the power distribution factor D j The weighted calculation above (t), the number of failures F k (t) in factor G l (t) above, and the operating factor H m (t)In reliability index I n Calculations above (t).

[0080] In step S11, the following mathematical model is used: Among them, α i , β i and γ i represents the parameters that control different stages, the parameter δ represents the key exploration-exploitation trade-off factor, and j It represents the hybrid exploration-exploitation strategy. This formula calculates the index Φ Grid , used to evaluate the effectiveness of vulnerability detection mechanisms;

[0081] In step S12, it is achieved through the following mathematical model: Among them, E represents the overall effectiveness of the entire system in the time period T, the parameters α and (1-α) represent the weights given to the exploration strategy and the utilization strategy respectively, and P explore (t) and P exploit (t) represents the corresponding probability or score at time t, the ratio represents a cost-effectiveness measure, where C(t) and P(t) are the energy production cost and consumer price at time t, respectively, and the exponent β regulates the system's sensitivity to this cost-price ratio.

[0082] In step S12, the Markov model optimization method is as follows:

[0083] S121, enter from the start module, and set the initial parameters of simulated annealing and reinforcement learning in the initialization module;

[0084] S122, enter the exploration phase. In the exploration initialization module, simulated annealing is allowed to widely explore the solution space, generate diverse initial solutions, avoid local optimal solutions, and discover potential global optimal solutions. Pre-training is used in the exploration-exploitation trade-off parameter module to determine the acceptance probability and exploration intensity of simulated annealing and the balance between exploring new states and exploiting existing knowledge in reinforcement learning.

[0085] S123. Utilize the optimization module to refine the solutions found in the exploration phase using reinforcement learning and focus on promising areas in the search space to improve the quality of the solution and the speed of convergence.

[0086] S124, using an adaptive exploration-exploitation balance module to monitor the convergence rate and performance indicators and dynamically adjust the balance between exploration and exploitation;

[0087] S125, measuring the performance of the optimized solution in the solution quality evaluation module and evaluating the reliability and efficiency of smart grid monitoring;

[0088] S126. In the convergence check module, ensure whether the optimization process meets the convergence criteria. If the convergence criteria are met, enter the optimal solution finding module, the algorithm outputs the optimized solution and ends the optimization process. If the convergence criteria are not met, the algorithm will check whether the iteration upper limit module is reached. If the iteration upper limit has been reached, the optimization process is terminated; if the iteration upper limit has not been reached, the algorithm returns to the exploration stage and continues the optimization.

[0089] S127. Throughout the entire process, the algorithm dynamically adjusts the exploration and development work in the module based on the progress and convergence indicators. At this time, the parameter update module adaptively adjusts the simulated annealing scheduling and reinforcement learning parameters to further optimize the exploration-exploitation process.

[0090] S128. Then, in the continuous optimization module of exploration-exploitation trade-off, the exploration-exploitation trade-off parameters are further refined to ensure the balance of the method during the optimization process.

[0091] S129. A hybrid simulated annealing-reinforcement learning approach is used in the hybrid strategy execution module: simulated annealing is responsible for exploration and generating diverse solutions; reinforcement learning is responsible for utilization and optimization of the solutions and convergence to the optimal solution. The optimization process is completed in the end module.

[0092] Markov model is used to optimize smart grid monitoring and has achieved significant technical results in many aspects. First, the model can optimize energy supply scheduling and reduce losses caused by energy non-supply (ENS). For example, the exploration initialization model provides subsidies when ENS is low, but as ENS increases, it gradually turns into economic losses, while other optimization models can better adapt to ENS changes, thereby reducing losses. This trend has been Figure 2 This is reflected in the . Furthermore, the model demonstrates strong adaptability in terms of grid topology and load management. Different models are suitable for grids of different sizes. The exploration-initialization model is suitable for small-scale grids, while the adaptive exploration-exploitation model can be used for larger grids. Furthermore, the improved load variation management capability enables the grid to stably cope with fluctuations from 15kW to 1000kW. See Table 1 for details.

[0093] Table 1 Smart grid transition scenarios and the application of the proposed Markov model in distribution automation

[0094]

[0095] In terms of power grid fault detection and service restoration, the Markov model improves fault detection speed and restoration efficiency, including the rapid identification and repair of voltage drops, current spikes, frequency fluctuations, and grid distortion, thereby improving grid stability. Figure 3

[0096] The model also optimizes outage management and improves grid reliability. In particular, the hybrid exploration-exploitation model maintains 100% customer satisfaction even after 80 outages, demonstrating that smart grids are more stable and adaptable in outage management. Figure 4 In addition, in terms of power grid security, the model performs well in vulnerability detection. The hybrid exploration-exploitation model has a detection rate of up to 90% at security scan level 1, effectively reducing the security risks faced by the power grid. Figure 5 Verified in. In terms of energy production and cost management, the model optimizes the balance between energy production costs and user electricity prices. The hybrid exploration-utilization model can maintain low consumer electricity prices when production costs are high, allowing power grid companies to provide more economical electricity while ensuring profitability. Figure 6In addition, by comparing with existing methods, the Markov model shows advantages in multiple key indicators, including reducing energy consumption by 7.5%, increasing renewable energy penetration by 15.5%, reducing power outage recovery time by 62.1%, reducing the number of successful cyber attacks by 32.6%, and improving market efficiency by 8.9%. The relevant comparative data are shown in Table 2.

[0097] Table 2 Comparison of existing methods and the proposed Markov model in distribution automation smart grid monitoring

[0098]

[0099] These results demonstrate that the Markov model not only improves grid efficiency but also enhances security and market adaptability, making it a valuable tool for smart grid optimization. In the future, the model could be combined with artificial intelligence and machine learning technologies to further enhance its predictive capabilities and intelligence, promoting the automated management and sustainable development of smart grids.

[0100] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and description of the present invention.

Claims

1. A Markov modeling method for automated smart grid monitoring, characterized in that: The following steps are involved: S1. Describe the probabilistic transition relationship of states over time in smart grid monitoring based on distributed automation; S2. Comprehensive metrics for evaluating Markov models for smart grid monitoring and distributed automation; S3, represents the state transition probability of the distribution automation system; S4, describing the probability transfer relationship of smart grid load changes; S5. Describe the state φ of the distribution automation system from time t t Transition to state φ at time t+1 t+1 probability; S6. Determine the daily load demand percentage; S7, describing the dynamic change of the state of charge of the kth energy storage unit in the smart grid; S8. Establish a fault detection and service restoration model in smart grid monitoring; S9, calculating the power quality index at time t; S10. Evaluate the impact of power outages and customer satisfaction; S11. Describe the comprehensive assessment of smart grid security using Markov models. S12. Evaluate smart grid monitoring system based on Markov model and distribution automation.

2. A Markov modeling method for automated smart grid monitoring according to claim 1, characterized in that: In the step S1, it is achieved by the following mathematical model: The future state depends only on the current state, not on the previous state sequence. In this model, t represents the state of the smart grid system at time t, where ξ t Represents a specific state in the state set S. This equation means that if the system is currently in state ξ t , then it transfers to state ξ at time t+1 t+1 The probability of the transition is the probability of all possible previous states ξ t-1 ∈S and sum the results, and press t-1 Transfer to ξ t , and then from ξ t Transfer to ξ t+1 The joint probability of is weighted; In the step S2, it is achieved by the following mathematical model: Where Φ represents an aggregated measure that combines the impact of energy reliability and financial consequences within a specific time period T. The integral is from 0 to T, capturing the cumulative effect of dynamic factors that affect energy supply reliability and related financial costs. In this equation, λ(t) and μ(t) are weighting functions used to adjust the unsupplied energy (E NS (t)) and financial impact (C F (t)) importance, E total (t) represents the total energy demand or capacity at time t, and provides a reference for evaluating the proportion of unsupplied energy to total demand. max Acts as a normalizing factor to ensure that the financial impact is assessed relative to the maximum allowed threshold.

3. The Markov modeling method for automated smart grid monitoring according to claim 1, characterized in that: In the step S3, it is achieved by the following mathematical model: Indicates that the system changes from the state φ at time t t Transition to state φt at time t+1 +1 The possibility of, given the state configuration st, where Φ t Indicates the current state of the distribution automation system, which is affected by the variable D t Influence, variable S t Contains properties related to the system operating settings and environmental conditions. This equation is valid for all possible configurations D t Sum up, each configuration represents a different operating scenario that affects state transition, conditional probability P(Φ t+1 =φ t+1 ∣Φ t =φ t ,S t =s t ,D t =dt) quantifies the t In the case of transfer to φ t+1 probability; In the step S4, it is achieved by the following mathematical model: At time t, it is in state λ t Consider the load characteristics After that, the system transfers to state λ at time t+1 t+1 The probability of t Indicates the current load change state in the smart grid system, which is affected by L t Variable influence, L t Including peak load, average load, minimum load, load curve and load composition parameters, the equation sums all possible load profiles P. Each load profile represents a different operating configuration and affects the state transition. The conditional probability Quantify the system transfer to λ t+1 possibility; In the step S5, it is achieved by the following mathematical model: Consider the current state φ t and fault characteristics f t , which sums all possible failure scenarios R, which affect the probability P(Φ t+1 =φt+1|Φ t =φ t ,F t =f t ,R t =f t ) This equation quantifies the impact of different fault rates, types, and locations on system behavior and is crucial for optimizing fault management strategies in smart grids.

4. The Markov modeling method for automated smart grid monitoring according to claim 1, characterized in that: In the step S6, it is achieved by the following mathematical model: It is crucial for the effectiveness of the proposed Markov model in smart grid management, where P(t) represents the percentage of total load demand at time t, which is calculated by subtracting the energy demand E from all N Markov models. i The sum of (t) and the distribution automation factor D in all M models j (t) and multiplied by 100. This formula is the cornerstone of using Markov models to evaluate and optimize energy distribution strategies. These models are an important part of the field of smart grid monitoring. With their inherent ability to explore and exploit system dynamics, they promote adaptive and efficient operation modes. In practical applications, each E i (t) represents the energy demand of the i-th Markov model at a certain time, reflecting the dynamic fluctuation of energy usage pattern, while D j (t) represents the distribution automation factor of the jth Markov model, reflecting the degree of automation control and optimization in the energy distribution process. By summarizing the demands and factors in multiple models, this equation provides a global view of the overall load demand over time; In the step S7, it is achieved by the following mathematical model: Among them, S k (t) represents the SoC state at time t, which is updated based on the current state, the energy output E from the N Markov models i (t) and the distribution automation factor D from M models j (t), parameter α k , β k and γ k The equation determines the charging and discharging dynamics and efficiency of the energy storage unit. It illustrates how the energy storage unit can be combined with the Markov model to optimize grid stability and energy distribution efficiency using exploration-exploitation strategies.

5. The Markov modeling method for automated smart grid monitoring according to claim 1, characterized in that: In the step S8, the following mathematical model is used: ij (t) = α ij ·V ij (t)+β ij ·C ij (t)-γ ij ·F ij-1 (t)+δ ij ·D ij (t), where F ij (t) represents the probability of event i and Markov model j failing at time t, affected by voltage sag / swell (V ij (t)), current spike (C ij (t)), previous fault status (F ij-1 (t)) and the distortion percentage (D ij (t)). Parameter α ij , β ij , γ ij and δ ij The impact of these factors on fault probability and service restoration decisions is determined. The equation illustrates how Markov models can integrate multiple metrics to enhance distribution automation capabilities and maintain grid stability and reliability under different electrical disturbance conditions. In the step S9, it is achieved by the following mathematical model: Its measurement range covers 0 to 360 degrees simultaneously. The index combines the energy demand E i (t) In the distribution automation factor D j The weighted average value above (t), and the number of failures F n (t) In the impact factor G p The index provides a comprehensive measure of grid stability and efficiency based on the trade-off calculation above (t). Influenced by the exploration-exploitation strategies in the Markov model, it promotes real-time evaluation and optimization in smart grid monitoring and automation, which is crucial for maintaining reliable power services and minimizing interference.

6. The Markov modeling method for automated smart grid monitoring according to claim 1, characterized in that: In the step S10, the following mathematical model is used: Through the energy demand E i (t) In the power distribution factor D j The weighted calculation above (t), the number of failures F k (t) in factor G l (t) above, and the operating factor H m (t)In reliability index I n Calculations above (t). In the step S11, the following mathematical model is used: Among them, α i , β i and γ i represents the parameters that control different stages, the parameter δ represents the key exploration-exploitation trade-off factor, and j It represents the hybrid exploration-exploitation strategy. This formula calculates the index Φ Grid , used to evaluate the effectiveness of vulnerability detection mechanisms; In the step S12, it is achieved by the following mathematical model: Among them, E represents the overall effectiveness of the entire system in the time period T, the parameters α and (1-α) represent the weights given to the exploration strategy and the utilization strategy respectively, and P explore (t) and P exploit (t) represents the corresponding probability or score at time t, the ratio represents a cost-effectiveness measure, where C(t) and P(t) are the energy production cost and consumer price at time t, respectively, and the exponent β regulates the system's sensitivity to this cost-price ratio.

7. A Markov modeling method for automated smart grid monitoring according to claim 6, characterized in that: In step S12, the Markov model optimization method is as follows: S121, enter from the start module, and set the initial parameters of simulated annealing and reinforcement learning in the initialization module; S122, enter the exploration phase. In the exploration initialization module, simulated annealing is allowed to widely explore the solution space, generate diverse initial solutions, avoid local optimal solutions, and discover potential global optimal solutions. Pre-training is used in the exploration-exploitation trade-off parameter module to determine the acceptance probability and exploration intensity of simulated annealing and the balance between exploring new states and exploiting existing knowledge in reinforcement learning. S123. Utilize the optimization module to refine the solutions found in the exploration phase using reinforcement learning and focus on promising areas in the search space to improve the quality of the solution and the speed of convergence. S124, using an adaptive exploration-exploitation balance module to monitor the convergence rate and performance indicators and dynamically adjust the balance between exploration and exploitation; S125, measuring the performance of the optimized solution in the solution quality evaluation module and evaluating the reliability and efficiency of smart grid monitoring; S126. In the convergence check module, ensure whether the optimization process meets the convergence criteria. If the convergence criteria are met, enter the optimal solution finding module, the algorithm outputs the optimized solution and ends the optimization process. If the convergence criteria are not met, the algorithm will check whether the iteration upper limit module is reached. If the iteration upper limit has been reached, the optimization process is terminated; if the iteration upper limit has not been reached, the algorithm returns to the exploration stage and continues the optimization. S127. Throughout the entire process, the algorithm dynamically adjusts the exploration and development work in the module based on the progress and convergence indicators. At this time, the parameter update module adaptively adjusts the simulated annealing scheduling and reinforcement learning parameters to further optimize the exploration-exploitation process. S128. Then, in the continuous optimization module of exploration-exploitation trade-off, the exploration-exploitation trade-off parameters are further refined to ensure the balance of the method during the optimization process. S129. A hybrid simulated annealing-reinforcement learning method is used in the hybrid strategy execution module: simulated annealing is responsible for exploration and generating diversified solutions; Reinforcement learning is responsible for utilizing,optimizing the solution and converging to the optimal one.,The optimization process is completed in the end module.

Images