A Sensor Scheduling Method and System Based on a Dual Threshold Structure

By adopting a sensor scheduling method based on a dual-threshold structure, the problems of poor communication stability and high cost in traditional power grid systems are solved, achieving efficient state estimation and communication optimization of smart grids, and reducing system complexity and cost.

CN117154924BActive Publication Date: 2026-11-17NORTHEASTERN UNIV CHINA
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202310855060.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-07-12
Publication Date
2026-11-17
Estimated Expiration
2043-07-12

AI Technical Summary

Technical Problem

Traditional power grid systems suffer from poor communication stability, limited communication resources, and high communication costs. This is especially true in smart grids, where the increased number of sensors and expanded coverage lead to increased system wiring complexity and design costs. Furthermore, the uncertainty of sensor energy results in poor data transmission stability.

Method used

A sensor scheduling method based on a dual-threshold structure is adopted. By modeling the power system as a discrete time-invariant system, multiple mathematical models are constructed and objective functions are generated. A Markov framework is constructed to simplify the objective function and determine the sensor transmission strategy, including a threshold structure based on the number of consecutive packet losses and battery capacity.

Benefits of technology

It improves the accuracy and stability of sensor transmission strategies, reduces communication frequency and cost, supports real-time monitoring and control of smart grids, and simplifies computational complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117154924B_ABST
    Figure CN117154924B_ABST
Patent Text Reader

Abstract

This invention relates to the field of communication technology and discloses a sensor scheduling method and system based on a dual-threshold structure. The method includes first modeling the power system as a discrete time-invariant system to obtain a power system model; then, based on the communication links between sensors and remote controllers in the power system model, constructing multiple mathematical models and calculating communication resource parameters based on these models; next, generating an objective function based on the communication resource parameters and constructing a Markov framework to simplify the objective function; finally, under preset conditions, determining a sensor transmission strategy with a dual-threshold structure based on the objective function. The dual-threshold structure includes a threshold structure based on the number of consecutive packet losses and a threshold structure based on battery capacity. This method can not only accurately predict the state of the power system but also effectively reduce communication frequency and save communication costs, bringing great convenience to the field of smart grids and possessing broad application prospects.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a sensor scheduling method and system based on a dual-threshold structure. Background Technology

[0002] With the advancement of modern technology, the problems of resource consumption and energy shortages have become increasingly apparent. Therefore, accelerating the development of smart grids powered primarily by renewable energy is a feasible solution to the global energy crisis. However, due to unpredictable weather conditions, the state (voltage, current) of smart grids also exhibits uncertainty. Consequently, a distribution management system is needed for high-level control and management of the distribution system. Among these systems, state estimation, as a crucial component of the distribution management system, plays a vital role in the monitoring and control of smart grids.

[0003] In traditional power grid systems, distribution management systems can only collect and monitor data in local areas to predict the state of the power grid. This cannot fully meet the real-time monitoring needs of large-scale smart grids. Furthermore, as the number of sensors increases and the coverage area expands, using dedicated independent connection methods increases the complexity of system wiring and design costs. At the same time, since the sensors are powered by batteries, the probability of successful data transmission is limited by the energy of the sensors. In addition, the energy collected by the sensors from the external environment is uncertain. Therefore, the communication process in traditional power grid systems suffers from problems such as poor stability, limited communication resources, and high communication costs. Summary of the Invention

[0004] In view of this, this application provides a sensor scheduling method and system based on a dual threshold structure, the main purpose of which is to solve the problems of poor stability, limited communication resources and high communication costs in the communication process of traditional power grid systems in the prior art.

[0005] According to a first aspect of the present invention, a sensor scheduling method based on a dual-threshold structure is provided, the method comprising:

[0006] The power system is modeled as a discrete time-invariant system, resulting in a power system model;

[0007] Based on the communication link between the sensor and the remote controller in the power system model, multiple mathematical models are constructed, and communication resource parameters are calculated based on the multiple mathematical models.

[0008] An objective function is generated based on the communication resource parameters, and a Markov framework is constructed. The objective function is simplified within the Markov framework, wherein the objective function is used to indicate the error covariance and communication cost of the remote controller.

[0009] Under preset conditions, a sensor transmission strategy with a dual-threshold structure is determined based on the objective function, wherein the dual-threshold structure includes a threshold structure based on the number of consecutive packet losses and a threshold structure based on battery capacity.

[0010] Optionally, the step of modeling the power system as a discrete time-invariant system to obtain a power system model includes:

[0011] Obtain the voltage values ​​of multiple distributed power sources in a power system;

[0012] Based on the voltage values ​​of the multiple distributed power sources, node voltage equations are constructed in the Laplace transform domain;

[0013] The node voltage equations are rewritten and transformed from the Laplace transform domain to the time domain to model the power system as a discrete time-invariant system, thus obtaining the power system model.

[0014] Optionally, based on the communication link between the sensor and the remote controller in the power system model, multiple mathematical models are constructed, and communication resource parameters are calculated according to the multiple mathematical models, including:

[0015] The communication links between the sensor and the remote controller in the power system model are obtained, and the packet loss links between the sensor, the remote controller, and the sensor and the remote controller that depend on the action are determined, wherein the sensor has an energy harvesting module;

[0016] A sensor model is constructed based on the sensor, and the first communication parameters are calculated based on the sensor model.

[0017] A channel model is constructed based on the packet loss link dependent on the aforementioned action, and the second communication parameters are calculated based on the channel model.

[0018] A controller model is constructed based on the remote controller, and the third communication parameter is calculated based on the controller model;

[0019] The first communication parameter, the second communication parameter, and the third communication parameter are integrated into a communication resource parameter.

[0020] Optionally, the step of constructing a sensor model based on the sensor and calculating the first communication parameter according to the sensor model includes:

[0021] Based on the power system model, the state estimation definition equation and the error covariance definition equation of the sensor are obtained;

[0022] Based on the Kalman filter, the state estimation definition equation and the error covariance definition equation are updated to obtain the state estimation calculation equation and the error covariance calculation equation of the sensor, and the sensor model is determined.

[0023] Based on the sensor model, the first communication parameters are obtained, wherein the first communication parameters include the sensor's own energy, the energy collected by the sensor, and the transfer probability of energy collection by the sensor in adjacent time moments.

[0024] Optionally, the step of constructing a channel model based on the packet loss link dependent on the action, and calculating the second communication parameters according to the channel model, includes:

[0025] In the power system model, packet loss links for dependent actions between the sensor and the remote controller are identified;

[0026] A channel model is constructed based on the packet loss link dependent on the action, and a second communication parameter is calculated based on the channel model. The second communication parameter includes the channel gain of the packet loss link, and the channel gain is used to indicate the decision state of the sensor.

[0027] The state transition probability of the sensor is determined based on the channel gain, wherein the state transition probability is determined by the signal-to-noise ratio of the sensor.

[0028] Optionally, the step of constructing a controller model based on the remote controller and calculating the third communication parameter according to the controller model includes:

[0029] A controller model is constructed based on the remote controller. In the controller model, the packet receiving indication variable of the remote controller is determined according to the state of the communication link and the decision state of the sensor, so as to obtain the probability of the remote controller receiving data packets.

[0030] The state estimation equation and error covariance equation of the remote controller are calculated based on the probability of the remote controller receiving data packets.

[0031] The time elapsed from the last successful transmission to the current moment is obtained by acquiring the remote controller, and the time is iterated to obtain the iteration time, which is used as the third communication parameter.

[0032] Optionally, constructing a Markov framework and simplifying the objective function within the Markov framework includes:

[0033] The state space of the remote controller is determined based on the communication resource parameters;

[0034] Obtain the action space of the remote controller, and determine the state transition probability of the remote controller based on the state space and the action space;

[0035] A one-step cost function is defined based on the state space and the action space, and the objective function is simplified based on the one-step cost function.

[0036] Optionally, the preset condition includes the objective function being at its minimum value; the step of determining the sensor transmission strategy with a dual-threshold structure based on the objective function under the preset condition includes:

[0037] A first calculation model and a second calculation model are established based on the preset conditions, respectively.

[0038] Based on the first calculation model, a threshold structure based on the number of consecutive packet losses is established, and a first transmission strategy is generated based on the threshold structure of the number of consecutive packet losses.

[0039] Based on the second calculation model, a threshold structure based on battery capacity is established, and a second transmission strategy is generated based on the threshold structure of consecutive packet loss.

[0040] The first transmission strategy and the second transmission strategy are integrated to obtain a sensor transmission strategy with a dual threshold structure.

[0041] Optionally, after determining the sensor transmission strategy with a dual-threshold structure based on the objective function, the method further includes:

[0042] Based on the sensor transmission strategy with the dual threshold structure, the determination conditions for maintaining the stability of the remote controller are obtained.

[0043] The sensor transmission strategy of the dual-threshold structure is determined based on the determination criteria to obtain the determination result.

[0044] When the determination result indicates that the sensor transmission strategy of the dual-threshold structure meets the determination condition, it is determined that the sensor transmission strategy of the dual-threshold structure is feasible.

[0045] According to a second aspect of the present invention, a sensor scheduling system based on a dual-threshold structure is provided, the system comprising:

[0046] The first modeling unit is used to model the power system as a discrete time-invariant system to obtain the power system model.

[0047] The second modeling unit is used to construct multiple mathematical models based on the communication link between the sensor and the remote controller in the power system model, and to calculate the communication resource parameters based on the multiple mathematical models.

[0048] A function generation unit is used to generate an objective function based on the communication resource parameters and construct a Markov framework to simplify the objective function within the Markov framework, wherein the objective function is used to indicate the error covariance and communication cost of the remote controller;

[0049] The strategy acquisition unit is used to determine a sensor transmission strategy with a dual threshold structure based on the objective function under preset conditions, wherein the dual threshold structure includes a threshold structure based on the number of consecutive packet losses and a threshold structure based on battery capacity.

[0050] The sensor scheduling method and system based on a dual-threshold structure provided by this invention considers the instability of energy acquisition from the external environment, which leads to the randomness of the actual energy of the sensor, and the relationship between the transmitted energy of the sensor and the packet reception rate at the controller. Therefore, the randomness of energy harvesting is coupled with the estimation error of the remote controller. A Markov framework is then constructed to model the external energy, simplifying the modeling process and providing tools for state transition probabilities and decision optimization, supporting the prediction and planning of scheduling methods. Simultaneously, this application considers that the state evolution of the communication channel depends on the transmission decision at that moment, aiming to minimize the estimation error at the controller and communication costs. The sensor transmission strategy obtained based on this objective has the property of a dual-threshold structure, reducing computational complexity. The above method can not only accurately predict the state of the power system but also effectively reduce communication frequency and save communication costs, bringing great convenience to the field of smart grids and possessing broad application prospects.

[0051] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0052] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0053] Figure 1 A schematic flowchart of a sensor scheduling method based on a dual-threshold structure provided by an embodiment of the present invention is shown;

[0054] Figure 2 The figure shows the uncontrolled system state during the simulation analysis of a sensor scheduling method based on a dual-threshold structure provided by an embodiment of the present invention.

[0055] Figure 3The figure shows a controlled system state diagram during the simulation analysis of a sensor scheduling method based on a dual-threshold structure provided by an embodiment of the present invention.

[0056] Figure 4 This paper presents a comparison chart of error covariance with and without a controller during the simulation analysis of a sensor scheduling method based on a dual-threshold structure provided by an embodiment of the present invention.

[0057] Figure 5 This diagram illustrates the system state estimation and actual values ​​under a sensor transmission strategy based on a dual-threshold structure during the simulation analysis of a sensor scheduling method based on a dual-threshold structure provided by an embodiment of the present invention.

[0058] Figure 6 This diagram illustrates a performance comparison of multiple strategies during the simulation analysis of a sensor scheduling method based on a dual-threshold structure, as provided in an embodiment of the present invention. Detailed Implementation

[0059] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0060] This application provides a sensor scheduling method based on a dual-threshold structure. The method specifically includes the following steps: First, the power system is modeled as a discrete time-invariant system to obtain a power system model. Then, based on the communication link between the sensor and the remote controller in the power system model, multiple mathematical models are constructed, and communication resource parameters are calculated according to the multiple mathematical models. Next, an objective function is generated based on the communication resource parameters, and a Markov framework is constructed to simplify the objective function within the Markov framework. The objective function is used to indicate the error covariance and communication cost of the remote controller. Finally, under preset conditions, a sensor transmission strategy based on the dual-threshold structure is determined based on the objective function. The dual-threshold structure includes a threshold structure based on the number of consecutive packet losses and a threshold structure based on battery capacity.

[0061] This invention provides a sensor scheduling method based on a dual-threshold structure. Considering the instability of energy acquisition from the external environment, which leads to the randomness of the sensor's actual energy, and the relationship between the sensor's transmitted energy and the packet reception rate at the controller, the randomness of energy harvesting is coupled with the estimation error of the remote controller. Therefore, a Markov framework is constructed to model the external energy, simplifying the modeling process and providing tools for state transition probabilities and decision optimization, supporting the prediction and planning of scheduling methods. Simultaneously, this application considers that the state evolution of the communication channel depends on the transmission decision at that moment, aiming to minimize the estimation error at the controller and communication costs. The sensor transmission strategy obtained based on this objective has the property of a dual-threshold structure, reducing computational complexity. This method can not only accurately predict the state of the power system but also effectively reduce communication frequency and save communication costs, bringing great convenience to the smart grid field and possessing broad application prospects.

[0062] Specifically, the power system is modeled as a discrete time-invariant system to obtain the power system model. This includes: first, obtaining the voltage values ​​of multiple distributed sources in the power system; then, based on the voltage values ​​of the multiple distributed sources, constructing nodal voltage equations in the Laplace transform domain; finally, rewriting the nodal voltage equations and converting them from the Laplace transform domain to the time domain to model the power system as a discrete time-invariant system and obtain the power system model.

[0063] In the embodiments provided in this application, N distributed power sources are first pre-configured as voltage sources and connected to the power system, with voltage values ​​as follows: Then, in the Laplace transform domain, the nodal voltage equation is:

[0064]

[0065] Where Y is the admittance matrix of the power system, {v kn (s)} n=1,2,...,N Let {L} be the voltage at the common coupling point. cn (s)} n=1,2,...,N It is a coupled inductor.

[0066] By defining v k =(v k1 ,...,v kN ) T v c =(v c1 ,...,v cN ) T and L = diag(L) c1 ,...,L cN Rewriting the node voltages yields:

[0067]

[0068] Transforming the Laplace transform domain back to the time domain, the power system is modeled as a discrete time-invariant system, as shown below:

[0069] x k+1 =Ax k +Bu k +ω k ,

[0070] y k =Cx k +v k .

[0071] Wherein, state x k ∈R M×1 Defined as an M-dimensional vector representing the voltage deviation from the reference value (the difference between the reference voltage and the voltage measured by distributed voltage sensors in the power system), the sensors measure the system state. Defined as n measured by the sensor c A state vector of dimension ω. A, B, and C are all system parameter matrices. Furthermore, due to the weak coupling between system dynamics and state evolution and control, the off-diagonal elements of system matrices A and B are not zero. k and v k These are Gaussian noises with a mean of 0 and covariance matrices W and V, respectively. k This represents the voltage control behavior performed by renewable energy sources. Furthermore, assuming the initial state x0 also follows Gaussian noise with a mean of 0 and a covariance of X0, then at all times k, assuming... The three are independent of each other.

[0072] Specifically, based on the communication link between the sensor and the remote controller in the power system model, multiple mathematical models are constructed, and communication resource parameters are calculated based on these models. This includes: first, obtaining the communication link between the sensor and the remote controller in the power system model, identifying the sensor, the remote controller, and the action-dependent packet loss link between the sensor and the remote controller (where the sensor has an energy harvesting module); then, constructing a sensor model based on the sensor, and calculating the first communication parameter based on the sensor model; next, constructing a channel model based on the action-dependent packet loss link, and calculating the second communication parameter based on the channel model; then, constructing a controller model based on the remote controller, and calculating the third communication parameter based on the controller model; finally, integrating the first, second, and third communication parameters into communication resource parameters.

[0073] In the embodiments provided in this application, the communication link between the sensor and the remote controller is determined in the power system model. Then, the sensor, the remote controller, and the packet loss link between the sensor and the remote controller that depends on the action are modeled and analyzed separately to ensure a comprehensive understanding of the communication status between the sensor and the remote controller in the smart grid, so as to achieve the purpose of comprehensive analysis and prediction.

[0074] Specifically, a sensor model is constructed based on the sensor, and the first communication parameters are calculated based on the sensor model. This includes: firstly, based on the power system model, obtaining the state estimation definition equation and the error covariance definition equation of the sensor; then, based on the Kalman filter, updating the state estimation definition equation and the error covariance definition equation to obtain the state estimation calculation equation and the error covariance calculation equation of the sensor, determining the sensor model; and finally, obtaining the first communication parameters based on the sensor model. The first communication parameters include the sensor's own energy, the energy collected by the sensor, and the transition probability of energy collection between adjacent time points.

[0075] In the embodiments provided in this application, the sensor, based on the measured values ​​and control inputs of the observed power system, obtains the defining equations for its state estimation and error covariance as follows:

[0076]

[0077]

[0078] Then, based on the standard Kalman filter, the result is updated to obtain... and The calculation formula is as follows:

[0079]

[0080]

[0081]

[0082]

[0083]

[0084] Then, by defining the Lyapunov and Ricardi operators, they are as follows:

[0085]

[0086]

[0087] The error covariance P on the sensor side k s It converges to a steady value at an exponential rate. and yes To simplify the analysis, we assume that the sensor has reached a steady-state value at the initial moment, that is, for all...

[0088] Since the sensor consumes energy to transmit data, this application considers a sensor equipped with a data collection device that can collect energy from the external environment. At time k, the energy collected by the sensor from the outside is H. k Assume H k It follows a Markov process, and its value range is: [H1,...,H]. max The transition probability of energy harvesting between adjacent time steps is: P(H) k+1 |H k )=α H Therefore, starting at time k+1, the energy B of the sensor... k+1 for:

[0089] B k+1 =min{B k +H k -a k Λ,B max}

[0090] Where a k This represents the action taken by the sensor at time k, Λ represents the energy consumed by the sensor to transmit data, and B... max This is the maximum energy that the sensor can store.

[0091] Specifically, a channel model is constructed based on the packet loss link dependent on the action, and a second communication parameter is calculated based on the channel model. This includes: first, determining the packet loss link dependent on the action between the sensor and the remote controller in the power system model; then, constructing a channel model based on the packet loss link dependent on the action, and calculating the second communication parameter based on the channel model. The second communication parameter includes the channel gain of the packet loss link, which is used to indicate the decision state of the sensor; and finally, determining the state transition probability of the sensor based on the channel gain, where the state transition probability is determined by the signal-to-noise ratio of the sensor.

[0092] In the embodiments provided in this application, the data is specifically transmitted to the controller via an action-dependent packet loss link, and the channel gain of this link is expressed as: n k The specific value is: And n k This is called a sensor-dependent decision-making state, emphasizing its dependence on current and past transmission actions. The specific transition probabilities are:

[0093]

[0094] Based on the above formula, it can be concluded that the channel state is poor when the sensor is active, and improves when the sensor is silent. If both types of actions transition to the same channel state, the transition probability under sensor silence is greater than the transition probability under sensor action.

[0095] Specific assumptions P(n) k+1 =n j |n k =n i ,a k =0)≥P(n) k+1 =n j |n k =n i ,a k =1), then the probability that the remote controller successfully receives the data packet at time k is determined by the signal-to-noise ratio at time k, that is:

[0096] P(γ k =1)=f(SNR) k )

[0097] in, σ 2 For channel noise, p k This refers to the transmission power required for sensor data transmission. It is a monotonically increasing continuous function.

[0098] Specifically, a controller model is constructed based on the remote controller, and a third communication parameter is calculated based on the controller model. This includes: first, a controller model is constructed based on the remote controller; in the controller model, the packet receiving indicator variable of the remote controller is determined according to the state of the communication link and the decision state of the sensor to obtain the probability of the remote controller receiving data packets; then, the state estimation equation and error covariance equation of the remote controller are calculated based on the probability of the remote controller receiving data packets; then, the time elapsed from the most recent successful transmission to the current time is obtained, and the time is iterated to obtain the iteration time, which is used as the third communication parameter.

[0099] In the embodiments provided in this application, γ is defined k γ is defined as the packet receiving indicator variable for the remote controller, which is related to the state of the communication channel and transmission decisions. k =1 indicates that data transmitted from the sensor was received at time k, which means the probability of receiving the data packet is expressed as:

[0100]

[0101] By definition This indicates the time elapsed since the last successful transmission by the remote controller.

[0102] τ k The iterative formula is as follows:

[0103]

[0104] Finally, the data packet containing the voltage deviation information is sent to the remote controller via the wireless network, and the controller calculates its state estimate based on the received information. And error covariance P k|k They are respectively:

[0105]

[0106]

[0107] This application aims to minimize the estimation error covariance and communication cost at the controller end within the infinite time domain, namely:

[0108]

[0109] Where β∈(0,1) is the discount coefficient, α is the balance factor, a larger α indicates that the objective function focuses on transmission cost, and a smaller α indicates that the objective function focuses on minimizing the error covariance. R>0 is the weight matrix.

[0110] The above optimization objective is an LQG problem, and its optimal cost is expressed as:

[0111]

[0112] The optimization objective is transformed into:

[0113]

[0114] Among them G ∞ =A T S ∞ A+WS ∞ S ∞ =A T S ∞ A+WA T S ∞ B(B T S ∞ B+R) -1 B N B T S ∞ A.

[0115] Specifically, a Markov framework is constructed, and the objective function is simplified within the Markov framework. This includes: first, determining the state space of the remote controller based on communication resource parameters; then, obtaining the action space of the remote controller; determining the state transition probabilities of the remote controller based on the state space and action space; and finally, defining a one-step cost function based on the state space and action space, and generating the objective function based on the one-step cost function.

[0116] In the embodiments provided in this application, a Markov framework is established, and the state space, i.e., s, is first determined. k =(τ k ,n k B k H k The state space S contains τ k ,n k B k H k The set of all possible state combinations. Then determine the action space U = {0,1}. N Let represent the set of all possible scheduling actions, where a k ∈U, a k =1 indicates that the sensor transmits data to the controller, a k =0 indicates that the sensor remains silent and does not transmit data, thus yielding the state transition probability:

[0117] P(s'|s,a)=P(τ k ,n k+1 B k+1 H k+1 |τ k-1 ,n k B k H k ,a k )P(n k+1 |n k ,a k )P(H k+1 |H k ,a k )

[0118] in:

[0119]

[0120]

[0121] P(H k+1 |H k )=α H

[0122] This leads to the first-step cost function.

[0123]

[0124] Among them, G k =A T S k+1 A+WS k S k =A T S k+1 A+WA T S k+1 B(B T S k+1 B+R) -1 B N B T S k+1 A

[0125] Considering the infinite time domain,

[0126] G ∞ =A T S ∞ A+WS ∞ S ∞ =A T S ∞ A+WA T S ∞ B(B T S ∞ B+R) -1 B N B T S ∞ A

[0127]

[0128] Therefore, the objective function becomes:

[0129] Suppose there exists a deterministic stationary policy such that the error covariance at the remote control is bounded, i.e., the following Bellman equation has a solution:

[0130]

[0131] Where s + V represents the next state after taking action a, given the current state as s. * This represents the optimal value function.

[0132] Specifically, the preset conditions include: the objective function is minimized; under the preset conditions, a sensor transmission strategy with a dual-threshold structure is determined based on the objective function, including: firstly, establishing a first calculation model and a second calculation model based on the preset conditions; then, establishing a threshold structure based on the number of consecutive packet losses according to the first calculation model, and generating a first transmission strategy based on the threshold structure of the number of consecutive packet losses; then, establishing a threshold structure based on battery capacity according to the second calculation model, and generating a second transmission strategy based on the threshold structure of the number of consecutive packet losses; finally, integrating the first transmission strategy and the second transmission strategy to obtain a sensor transmission strategy with a dual-threshold structure.

[0133] In the embodiments provided in this application, an optimal strategy is assumed to exist. The present invention first provides a strategy based on a threshold structure of consecutive packet loss.

[0134] 1. Explanation:

[0135] For any k-time, the state s = (s 1,k ,s 2,k ,s 3,k ,s 4,k )∈S,

[0136] Define s 1,k =τ k-1 ,s 2,k =B k ,s 3,k =n k ,s 4,k =H k S1 = Z + ,

[0137] They belong to S2 = [0, B] respectively. max S3 = [1, n] s It can be seen that the state spaces S1, S2, S3, and S4 are all countable.

[0138] 2. Step One:

[0139] because Since s1 is non-decreasing, c(s,a) is non-decreasing with respect to s1 for all a∈U.

[0140] Furthermore, for a fixed s1∈S1, s3∈S3, s4∈S4, c(s,a) is valid for all s on S1×U. and a + ≥a - This is an additive function.

[0141] Therefore, for all a∈U and any fixed s2∈S2, s3∈S3, s4∈S4, c(s,a) is non-decreasing with respect to s1.

[0142] 3. Step Two:

[0143] Define two states as follows: S1(τ) + ,B1,n1,H1),S2(τ - (,B2,n2,H2),

[0144] For any fixed s'=(τ' k ,B' k+1 ,n' k+1 ,H' k+1 ), set τ +' ≥τ -' ≥τ' k+1 , B1=B2, n1=n2, H1=H2.

[0145] For any a k ∈U, when τ +' ≥τ -' ≥τ' k+1 At that time, we can obtain:

[0146]

[0147] When τ +' ≥τ' k+1 ≥τ -' hour:

[0148]

[0149] When τ' k+1 ≥τ + ≥τ - hour:

[0150]

[0151] Therefore: for a fixed s2∈S2, s3∈S3, s4∈S4, for all s'∈S, a∈U, Regarding τ k-1 It is not subtractive.

[0152] 4. Step Three:

[0153] When s 2,k ,s 3,k ,s 4,k It is fixed, for any fixed s' = (τ') on S. k ,B' k+1 ,n' k+1 ,H' k+1 ), on S1 U We can obtain:

[0154] When τ'k+1 ≥τ +' ≥τ -' hour:

[0155]

[0156] When τ +' ≥τ' k+1 ≥τ -' hour:

[0157]

[0158]

[0159] When τ +′ ≥τ -′ ≥τ′ k+1 hour:

[0160]

[0161]

[0162] Therefore, for all s'∈S and the fixed s2∈S2, s3∈S3, s4∈S4, It is a secondary additive function of s1 and a over S1×U.

[0163] From steps one to three, it can be seen that for any fixed n i ∈[1,N s ],τ∈Z + ,H∈[0,H max ], optimal strategy a * (s) The battery capacity B is non-decreasing. That is, it can be expressed as:

[0164] For an unstable system, the fixed channel state n i (i = 1, 2, ..., n) s The battery capacity B∈[0,B] max Collect energy H∈[0,H] max For all The transmission strategy is related to the number of consecutive packet losses τ. ni Non-subtractive, specifically in the form of:

[0165]

[0166] Furthermore, this application provides a threshold structure for the transmission strategy with respect to battery capacity.

[0167] 1. Explanation:

[0168] definition

[0169] s 1,k =τ k-1 ,s 2,k =B k ,s 3,k =n k ,

[0170] s 4,k =H k S1 = Z + S2=[0,B max S3 = [1, n] s S4 = [0, H] max ].

[0171] because With B k Since they are independent, c(s,a) is non-decreasing with respect to s² for all a∈U. For any fixed s'=(τ') on S, k ,B' k+1 ,n' k+1 ,H' k+1 ), fixed B k ,n k H k For S2 and any a k ∈U,

[0172] When B +' ≥B -' ≥B' k+1 hour:

[0173]

[0174] When B +' ≥B' k+1 ≥B -' hour:

[0175]

[0176] When B' k+1 ≥B +' ≥B -' hour:

[0177]

[0178] Therefore, for a fixed s1∈S1, s3∈S3, s4∈S4, all s'∈S, a∈U, Regarding B k-1 It is not subtractive.

[0179] When s 1,k ,s 3,k ,s 4,kIt is fixed, for any fixed s' = (τ') on S. k ,B' k+1 ,n' k+1 ,H' k+1 ), on S2 U

[0180] When B' k+1 ≥B +' ≥B -' hour:

[0181]

[0182] When B +' ≥B' k+1 ≥B -' hour:

[0183]

[0184]

[0185] When B +' ≥B -' ≥B' k+1 hour:

[0186]

[0187]

[0188] Therefore, for all s'∈S and the fixed s1∈S1, s3∈S3, s4∈S4, It is a secondary additive function of s2 and a over S2×U.

[0189] Based on the above, for any fixed n i ∈[1,N s ],τ∈Z + ,H∈[0,H max ], optimal strategy a * (s) The battery capacity B is non-decreasing. That is, it can be expressed as:

[0190] For an unstable system, the fixed channel state n i (i = 1, 2, ..., n) s ), consecutive packet loss number τ, energy collection H∈[0,H max For all The transmission strategy is about the number of consecutive packet losses B. * Non-subtractive, specifically in the form of:

[0191]

[0192] Specifically, after determining the sensor transmission strategy of the dual-threshold structure based on the objective function, the method further includes: firstly, obtaining the judgment conditions for maintaining the stability of the remote controller based on the sensor transmission strategy of the dual-threshold structure; then, judging the sensor transmission strategy of the dual-threshold structure based on the judgment conditions to obtain the judgment result; when the judgment result indicates that the sensor transmission strategy of the dual-threshold structure meets the judgment conditions, it is determined that the sensor transmission strategy of the dual-threshold structure is feasible.

[0193] In the embodiments provided in this application, in order to demonstrate that the above-described threshold type strategy can enable remote control

[0194]

[0195]

[0196] The estimation error covariance at the control end is bounded. First, we give the definition: a stationary deterministic strategy is determined by the channel state n. k Mapping of action a: N s →[0,1] determines the probability of attempting transmission based on action-dependent states, i.e.:

[0197] Among them, in n k =n j When k≥1, a k and They are independent of each other.

[0198] Then, under the given assumptions, we can show that a threshold-type strategy can make the system stable.

[0199] Assumption:

[0200]

[0201] in,

[0202]

[0203]

[0204] n j ,n j' ∈N s ,a∈U

[0205] Under the given threshold structure strategy, taking the threshold structure based on the number of consecutive packet losses as an example, if the current state is less than the threshold for the number of consecutive packet losses, then the remote estimator is bounded. If the current state is greater than the number of consecutive packet losses, then data is transmitted in the current state. If the transmission is successful, the remote controller receives the packet, and thus the estimator is bounded. If it fails, the error covariance on the remote controller side may become unbounded. Under unbounded covariance, the system has no optimal strategy. Under Assumption 1, the unbounded situation on the controller side will not occur. Therefore, two types of transmission strategies based on the threshold structure are given. * This ensures stability on the remote controller side, thereby satisfying the Bellman equation:

[0206]

[0207] In other words, the given threshold strategy is the optimal strategy, under which the remote controller can be made stable.

[0208] Specifically, in the embodiments provided in this application, the designed threshold-based transmission strategy provided in this application is applied to the voltage stabilization process of renewable energy in a smart grid for simulation verification. The method aims to optimize controller performance and communication costs. It utilizes the four-bus model of the distribution test feeder from the Distribution System Analysis Subcommittee's Distribution Test Feeder Working Group to model the state estimation model of the smart grid system, where voltage stabilization is achieved by renewable energy on multiple PCCs.

[0209] The system dynamic parameter matrices A and B are:

[0210]

[0211]

[0212] The system state represents the voltage deviation (the difference between the reference voltage value and the voltage measured by voltage sensors in the power system) on the four distributed energy generators, and the observation matrix C is the identity matrix. k This represents the multidimensional voltage control action performed by the distributed generation, whose optimal feedback gain matrix is ​​obtained by solving the LQR equation. The covariance matrices of the system state evolution and the Gaussian noise of the voltage deviation observations are respectively:

[0213] W=diag(0.01,0.01,0.01,0.01)×10 -2

[0214] V=diag(0.01,0.02,0.01,0.06)×10 -2

[0215] From the above system parameters, it can be seen that A has a non-negative eigenvalue; therefore, the system is unstable without feedback control. Based on the method provided in this application, it is first verified that under optimal control, the system state tends to stabilize around the voltage, and the error covariance of the system matrix reaches a steady state in a short time. When the optimal controller is not used, the estimated error covariance shows a gradual increasing trend after a period of time, and the system state is divergent, specifically as follows: Figures 2 to 4 As shown:

[0216] Figure 5 Given the four bus voltage states, the actual state of the system and the estimated state at the controller are obtained using the designed threshold-based state estimation method. To illustrate that this method can effectively control the dynamics of the power system, a comparison diagram of the actual state and the estimated state of each of the four buses is given.

[0217] To demonstrate the superiority of the method provided in this application, it is compared with well-known greedy strategies, random strategies, and Q-learning heuristic strategies. Figure 6 It can be seen that the threshold-type strategy provided in this application is relatively optimal, with the smallest trace of the average estimation error covariance and the best estimation performance.

[0218] The transmission strategy designed in this invention has threshold structures for consecutive packet loss and battery capacity, respectively. That is, when considering the threshold structure of the strategy with respect to consecutive packet loss, the battery capacity, channel state and collected energy are fixed, and the strategy will be in a state of continuous transmission when the consecutive packet loss is greater than the threshold, until the consecutive packet loss is less than the threshold structure.

[0219] This application also provides a sensor scheduling system based on a dual-threshold structure, specifically comprising: a first modeling unit for modeling the smart grid as a discrete time-invariant system to obtain a power system model; a second modeling unit for constructing multiple mathematical models based on the communication links between sensors and remote controllers in the power system model, and calculating communication resource parameters based on the multiple mathematical models; a function generation unit for constructing a Markov framework based on the communication resource parameters, and generating an objective function within the Markov framework, wherein the objective function is used to indicate the error covariance and communication cost of the remote controller; and a strategy acquisition unit for determining a sensor transmission strategy based on the dual-threshold structure under preset conditions, wherein the dual-threshold structure includes a threshold structure based on the number of consecutive packet losses and a threshold structure based on battery capacity.

[0220] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing this application. Those skilled in the art will understand that the modules in the apparatus of the embodiment can be distributed within the apparatus of the embodiment as described, or can be modified to be located in one or more apparatuses different from this embodiment. The modules of the above-described embodiment can be combined into one module, or further divided into multiple sub-modules.

[0221] The serial numbers in this application are for descriptive purposes only and do not represent the superiority or inferiority of any particular implementation scenario. The above disclosures are merely a few specific implementation scenarios of this application; however, this application is not limited thereto, and any variations conceived by those skilled in the art should fall within the protection scope of this application.

Claims

1. A sensor scheduling method based on a dual-threshold structure, characterized in that, The method includes: The power system is modeled as a discrete time-invariant system, resulting in a power system model; Based on the communication link between the sensor and the remote controller in the power system model, multiple mathematical models are constructed, and communication resource parameters are calculated based on the multiple mathematical models. An objective function is generated based on the communication resource parameters, and a Markov framework is constructed. The objective function is simplified within the Markov framework, wherein the objective function is used to indicate the error covariance and communication cost of the remote controller. Under preset conditions, a sensor transmission strategy with a dual threshold structure is determined based on the objective function, wherein the dual threshold structure includes a threshold structure based on the number of consecutive packet losses and a threshold structure based on battery capacity. The process of modeling the power system as a discrete time-invariant system to obtain a power system model includes: Obtain the voltage values ​​of multiple distributed power sources in a power system; Based on the voltage values ​​of the multiple distributed power sources, node voltage equations are constructed in the Laplace transform domain; The node voltage equations are rewritten and transformed from the Laplace transform domain to the time domain to model the power system as a discrete time-invariant system, thus obtaining the power system model. Based on the communication link between sensors and remote controllers in the power system model, multiple mathematical models are constructed, and communication resource parameters are calculated according to the multiple mathematical models, including: The communication links between the sensor and the remote controller in the power system model are obtained, and the packet loss links between the sensor, the remote controller, and the sensor and the remote controller that depend on the action are determined, wherein the sensor has an energy harvesting module; A sensor model is constructed based on the sensor, and the first communication parameters are calculated based on the sensor model. A channel model is constructed based on the packet loss link dependent on the aforementioned action, and the second communication parameters are calculated based on the channel model. A controller model is constructed based on the remote controller, and the third communication parameter is calculated based on the controller model; The first communication parameter, the second communication parameter, and the third communication parameter are integrated into a communication resource parameter; The preset conditions include the objective function being at its minimum value; the step of determining the sensor transmission strategy with a dual-threshold structure based on the objective function under the preset conditions includes: A first calculation model and a second calculation model are established based on the preset conditions, respectively. Based on the first calculation model, a threshold structure based on the number of consecutive packet losses is established, and a first transmission strategy is generated based on the threshold structure of the number of consecutive packet losses. Based on the second calculation model, a threshold structure based on battery capacity is established, and a second transmission strategy is generated based on the threshold structure of consecutive packet loss. The first transmission strategy and the second transmission strategy are integrated to obtain a sensor transmission strategy with a dual threshold structure.

2. The method according to claim 1, characterized in that, The step of constructing a sensor model based on the sensor and calculating the first communication parameter based on the sensor model includes: Based on the power system model, the state estimation definition equation and the error covariance definition equation of the sensor are obtained; Based on the Kalman filter, the state estimation definition equation and the error covariance definition equation are updated to obtain the state estimation calculation equation and the error covariance calculation equation of the sensor, and the sensor model is determined. Based on the sensor model, the first communication parameters are obtained, wherein the first communication parameters include the sensor's own energy, the energy collected by the sensor, and the transfer probability of energy collection by the sensor in adjacent time moments.

3. The method according to claim 1, characterized in that, The process of constructing a channel model based on the packet loss link dependent on the action, and calculating the second communication parameters based on the channel model, includes: In the power system model, packet loss links for dependent actions between the sensor and the remote controller are identified; A channel model is constructed based on the packet loss link dependent on the action, and a second communication parameter is calculated based on the channel model. The second communication parameter includes the channel gain of the packet loss link, and the channel gain is used to indicate the decision state of the sensor. The state transition probability of the sensor is determined based on the channel gain, wherein the state transition probability is determined by the signal-to-noise ratio of the sensor.

4. The method according to claim 1, characterized in that, The process of constructing a controller model based on the remote controller and calculating the third communication parameter based on the controller model includes: A controller model is constructed based on the remote controller. In the controller model, the packet receiving indication variable of the remote controller is determined according to the state of the communication link and the decision state of the sensor, so as to obtain the probability of the remote controller receiving data packets. The state estimation equation and error covariance equation of the remote controller are calculated based on the probability of the remote controller receiving data packets. The time elapsed from the last successful transmission to the current moment is obtained by acquiring the remote controller, and the time is iterated to obtain the iteration time, which is used as the third communication parameter.

5. The method according to claim 1, characterized in that, The construction of the Markov framework, within which the objective function is simplified, includes: The state space of the remote controller is determined based on the communication resource parameters; Obtain the action space of the remote controller, and determine the state transition probability of the remote controller based on the state space and the action space; A one-step cost function is defined based on the state space and the action space, and the objective function is simplified based on the one-step cost function.

6. The method according to claim 1, characterized in that, After determining the sensor transmission strategy with a dual-threshold structure based on the objective function, the method further includes: Based on the sensor transmission strategy with the dual threshold structure, the determination conditions for maintaining the stability of the remote controller are obtained. The sensor transmission strategy of the dual-threshold structure is determined based on the determination criteria to obtain the determination result. When the determination result indicates that the sensor transmission strategy of the dual-threshold structure meets the determination condition, it is determined that the sensor transmission strategy of the dual-threshold structure is feasible.

7. A sensor scheduling system based on a dual-threshold structure, characterized in that, The system includes: The first modeling unit is used to model the power system as a discrete time-invariant system to obtain the power system model. The second modeling unit is used to construct multiple mathematical models based on the communication link between the sensor and the remote controller in the power system model, and to calculate the communication resource parameters based on the multiple mathematical models. A function generation unit is used to generate an objective function based on the communication resource parameters and construct a Markov framework to generate the objective function within the Markov framework, wherein the objective function is used to indicate the error covariance and communication cost of the remote controller; The strategy acquisition unit is used to determine a sensor transmission strategy with a dual threshold structure based on the objective function under preset conditions, wherein the dual threshold structure includes a threshold structure based on the number of consecutive packet losses and a threshold structure based on battery capacity. The first modeling unit is also used to obtain the voltage values ​​of multiple distributed power sources in the power system; Based on the voltage values ​​of the multiple distributed power sources, node voltage equations are constructed in the Laplace transform domain; The node voltage equations are rewritten and transformed from the Laplace transform domain to the time domain to model the power system as a discrete time-invariant system, thus obtaining the power system model. The second modeling unit is also used to obtain the communication link between the sensor and the remote controller in the power system model, and to determine the packet loss link between the sensor, the remote controller, and the sensor and the remote controller that depends on the action, wherein the sensor has an energy harvesting module; A sensor model is constructed based on the sensor, and the first communication parameters are calculated based on the sensor model. A channel model is constructed based on the packet loss link dependent on the aforementioned action, and the second communication parameters are calculated based on the channel model. A controller model is constructed based on the remote controller, and the third communication parameter is calculated based on the controller model; The first communication parameter, the second communication parameter, and the third communication parameter are integrated into a communication resource parameter; The preset conditions include the objective function being the minimum value; the strategy acquisition unit is also used to establish a first calculation model and a second calculation model based on the preset conditions; Based on the first calculation model, a threshold structure based on the number of consecutive packet losses is established, and a first transmission strategy is generated based on the threshold structure of the number of consecutive packet losses. Based on the second calculation model, a threshold structure based on battery capacity is established, and a second transmission strategy is generated based on the threshold structure of consecutive packet loss. The first transmission strategy and the second transmission strategy are integrated to obtain a sensor transmission strategy with a dual threshold structure.

Citation Information

Patent Citations

  • Event trigger state estimating system based on markovian jumping

    CN109410361A

  • Remote state estimation sensor scheduling method, scheduler and information physical system

    CN115941802A