A smart grid energy efficiency analysis and optimization system

Abstract

The present invention relates to the field of smart grid energy efficiency analysis and optimization, and specifically to a smart grid energy efficiency analysis and optimization system, wherein a feature extraction module of the system is used to collect a set of smart grid operation status parameters and a set of communication QoS demand parameters and extract a smart grid operation status feature vector; a spectrum resource demand intensity coefficient sequence analysis module is used to obtain a reference spectrum resource demand intensity coefficient, and the reference spectrum resource demand intensity coefficient is weighted according to a service priority weight to obtain a spectrum resource demand intensity coefficient sequence; a spectrum resource allocation module is used to execute a spectrum resource allocation strategy according to the spectrum resource demand intensity coefficient sequence; and an energy efficiency optimization control module is used to optimize spectrum resource allocation through data analysis to minimize a communication delay-energy efficiency loss mapping function. The spectrum resources of the smart grid are ensured to be reasonably allocated, thereby reducing communication delay and energy efficiency loss while improving the energy efficiency of the smart grid.

Description

Technical Field

[0001] The present invention relates to the field of smart grid energy efficiency analysis and optimization, and in particular to a smart grid energy difference efficiency analysis and optimization system. Background Art

[0002] As the core form of the new generation of power systems, smart grids have achieved intelligent management of power production, transmission, distribution and consumption by integrating advanced information technology, communication technology and automation technology. Its core advantage lies in relying on real-time data transmission, such as sensor data, control instructions, etc. Therefore, the wireless communication system plays a key role in the neural network, supporting key businesses such as smart meter data return (AMI), distribution automation terminal control, and demand response instruction transmission.

[0003] However, the current smart grid wireless communication faces severe challenges of spectrum resource constraints: on the one hand, the power-specific frequency band has limited carrying capacity and is difficult to meet the exponentially growing access needs of massive smart terminals, especially with the rapid growth of smart grid equipment and applications; on the other hand, the public network frequency band has serious co-frequency interference problems, which not only leads to an imbalance in the supply and demand of spectrum resources, but also causes the transmission delay of key control instructions to fluctuate too much. Especially in the case of tight spectrum resources, the communication of various services in the smart grid will be congested and delayed, seriously affecting the real-time monitoring and control capabilities, and then causing the power grid to be unable to respond quickly to changing load demands, resulting in energy waste or inefficiency. For example, in the demand response (DR) system, the probability of failure to transmit control instructions during peak hours increases, forcing the power grid to call high-cost backup power supplies, which in turn leads to a decrease in energy utilization efficiency.

[0004] Therefore, there is an urgent need for a system that analyzes spectrum resources in smart grids and manages them efficiently to ensure the stability of communication delays and improve the real-time regulation capabilities of smart grids, so as to achieve energy efficiency optimization of smart grids. Summary of the invention

[0005] (1) Technical issues to be resolved

[0006] The present invention provides a smart grid energy efficiency analysis and optimization system to solve the problem of high service communication delay in the smart grid due to unreasonable spectrum resource allocation, which increases the energy efficiency loss of the grid.

[0007] (2) Technical solution

[0008] To achieve the above-mentioned object, the present invention provides a smart grid energy efficiency analysis and optimization system, characterized in that the system comprises: a feature extraction module, a spectrum resource demand intensity coefficient sequence analysis module, a spectrum resource allocation module, and an energy efficiency optimization control module, and each module is sequentially connected in communication;

[0009] The feature extraction module is used to collect the smart grid operation status parameter set and the communication QoS requirement parameter set in real time through the sensor network deployed at the intersection node of the smart grid physical layer and the communication layer, and extract the smart grid operation status feature vector after normalization through the sliding time window; the smart grid operation status parameter set includes voltage deviation and load mutation gradient, and the communication QoS requirement parameter set includes delay sensitivity, packet loss rate threshold, and service priority weight.

[0010] The spectrum resource demand intensity coefficient sequence analysis module is used to analyze the quantitative mapping relationship between the smart grid operation status and the communication QoS requirement through the spatiotemporal association rule base of the smart grid operation status feature vector to obtain the benchmark spectrum resource demand intensity coefficient; and to weight the benchmark spectrum resource demand intensity coefficient according to the service priority weight to obtain the spectrum resource demand intensity coefficient sequence.

[0011] A spectrum resource allocation module is used to execute a spectrum resource allocation strategy according to the spectrum resource demand intensity coefficient sequence; the spectrum resource allocation strategy includes allocating a hard slice layer to protection and control services and allocating a soft slice layer to metering and collection services; the spectrum resource allocation strategy is aligned with the timestamp of the smart grid control instruction through the time-sensitive network protocol, and it is ensured that the spectrum resource allocation is preloaded within a preset delay threshold before the control instruction is issued.

[0012] The energy efficiency optimization control module is used to obtain the communication delay of all services and calculate the delay standard deviation after the spectrum resource allocation strategy is executed, and construct a communication delay-energy efficiency loss mapping function by analyzing the delay standard deviation and energy efficiency index through data analysis; optimize the spectrum resource allocation through data analysis to minimize the communication delay-energy efficiency loss mapping function, and the delay standard deviation is less than the preset safety delay standard deviation threshold.

[0013] Furthermore, the spectrum resource demand intensity coefficient sequence analysis module includes:

[0014] The coupling feature vector generation module is used to construct the communication link-power equipment association matrix based on the physical connection relationship between power equipment and communication links in the smart grid topology. ,express Power equipment and The connection relationship matrix between the communication links; among them, when the control instruction of the ath power equipment depends on the When a communication link is transmitted, ,otherwise ; , .

[0015] The incidence matrix Perform weighted processing to obtain the weighted correlation matrix ; The weight value of the weighted processing ,in, For the The electrical centrality of an electrical device, For the The number of transmission hops of a communication link.

[0016] The weighted incidence matrix and the smart grid operation status feature vector Perform tensor fusion to generate coupled feature vectors ; The smart grid operation status feature vector ,in, is the power imbalance, is the time window for load mutation, is the load mutation gradient; is the voltage deviation; is the delay sensitivity; is the packet loss rate threshold; For Business The priority weight of .

[0017] Spectrum resource demand intensity coefficient sequence calculation submodule, used to obtain the spatiotemporal association rule base },in, for The coupled eigenvector at time instant; = for Communication quality requirements at all times, including measured latency , Packet loss rate and spectrum resource demand intensity coefficient benchmark value .

[0018] The spectrum resource demand intensity coefficient benchmark value is obtained by training the spatiotemporal correlation model through the long short-term memory network. As the target value of the spatiotemporal correlation model, the network input is , output the predicted spectrum resource demand intensity coefficient , the loss function is: ,in, is the transient stability margin of the smart grid, is the preset weight coefficient;

[0019] The coupled eigenvector Input the trained spatiotemporal correlation model and output the spectrum resource demand intensity coefficient .

[0020] Based on business priority weight Spectrum resource demand intensity coefficient Make corrections to generate spectrum resource demand intensity coefficient sequences for each service ,in, , For Business The spectrum resource demand intensity coefficient is It is the sum of the priority weights of all services.

[0021] Furthermore, the spectrum resource requirement intensity coefficient sequence calculation submodule includes the following steps:

[0022] The spectrum resource demand intensity benchmark value is obtained by using the spectrum resource demand intensity coefficient benchmark value calculation formula ; Spectrum resource demand intensity coefficient benchmark value The calculation formula is:

[0023] , and The rate of change is less than the preset Rate of change threshold.

[0024] in, , is the adaptive correction factor, , The initial value of is preset according to the smart grid topology; is the disturbance level factor of smart grid; is the response speed of the smart grid to disturbances, is the transient response time constant of the smart grid, The current timestamp.

[0025] The smart grid disturbance level factor The calculation formula is:

[0026] .

[0027] in, is the inertia time constant of the smart grid; is the power imbalance.

[0028] Furthermore, the spectrum resource allocation module includes:

[0029] The slice bandwidth management module is used for allocating the hard slice layer to the protection control service to allocate exclusive spectrum slices to the protection control service, and the hard slice bandwidth of the exclusive spectrum slice ,in, To protect bandwidth and meet ; is the total available spectrum bandwidth; For Business Hard slice bandwidth; It is the sum of the spectrum resource demand intensity coefficients of all services.

[0030] The allocation of soft slice layers to metering and collection services is to use a non-orthogonal multiple access method for metering and collection services, and the soft slice bandwidth of the non-orthogonal multiple access method is ;in, is the dynamic interference suppression factor, .

[0031] The hard slice bandwidth and soft slice bandwidth constraints satisfy ,and in, The preset minimum protection bandwidth.

[0032] Furthermore, the spectrum resource allocation module further includes the following steps:

[0033] The spectrum resource allocation strategy is aligned with the timestamp of the smart grid control instruction through the time-sensitive network protocol and a synchronization error is calculated, and the synchronization error is less than a preset synchronization error threshold.

[0034] Hard slice bandwidth preloading time window in spectrum resource allocation Transient response time constant of smart grid Satisfy constraints .

[0035] Furthermore, the energy efficiency optimization control module includes:

[0036] Energy efficiency analysis module, used for energy efficiency indicators including the reduction of grid backup capacity redundancy , Increase in renewable energy consumption rate ; The communication delay-energy efficiency loss mapping function Used to quantify the energy efficiency loss of the power grid according to the fluctuation of communication delay; communication delay-energy efficiency loss mapping function The calculation formula is:

[0037] .

[0038] in, is the standard deviation of the delay; is the normalized weight coefficient and satisfies .

[0039] Mapping function of communication delay-energy efficiency loss Input the reward function and dynamically adjust the spectrum allocation strategy to minimize it through the double-delayed deep deterministic policy gradient algorithm, and update the adaptive correction factor , .

[0040] Furthermore, the energy efficiency analysis module includes the following steps:

[0041] The reward function for:

[0042] .

[0043] in, Rewards for meeting the delay target; Penalty for exceeding the delay limit; Penalties for energy efficiency losses; reward for interference suppression; is the indicator function, when 1 if yes, 0 otherwise. is the preset safety delay standard deviation threshold; is a trainable hyperparameter that satisfies .

[0044] Furthermore, the energy efficiency analysis module also includes:

[0045] Adaptive correction factor update analysis module for adaptive correction factor Updates meet , adaptive correction factor Updates meet , ensuring the spectrum resource demand intensity benchmark value Related to energy efficiency indicators.

[0046] Furthermore, the slice bandwidth management module further includes:

[0047] The energy efficiency priority arbitration module is used to trigger energy efficiency priority arbitration when the hard slice bandwidth demand exceeds the total spectrum resources.

[0048] The energy efficiency priority arbitration is to obtain the energy efficiency impact factor of each business , according to the energy efficiency influencing factor Allocate hard slice bandwidth in descending order until ∑ ≤ ; The energy efficiency influencing factor The calculation formula is: ;in, For Business contribution of renewable energy consumption.

[0049] (3) Beneficial effects

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. By collecting the operating status parameters and communication QoS requirement parameters of the smart grid in real time, combined with the analysis and optimization of the spectrum resource demand intensity coefficient, the spectrum resources are reasonably allocated. The balance between communication delay and energy efficiency is optimized, and the energy efficiency loss in the process of reducing the redundancy of the grid's spare capacity and increasing the renewable energy consumption rate is reduced by adjusting and optimizing the spectrum allocation through adaptive correction factors.

[0052] 2. Through the slice bandwidth management module, the hard slice and soft slice bandwidth are reasonably allocated to ensure that the protection control services and metering collection services meet different service priorities and maximize network utilization efficiency. The dynamic interference suppression factor and energy efficiency priority arbitration mechanism effectively solve the problem of how to optimize resource allocation when spectrum resources are tight, improve the overall system performance while ensuring the achievement of communication latency and energy efficiency goals. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 This is a schematic diagram of the module composition of a smart grid energy efficiency analysis and optimization system according to Example 1 of the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Before giving examples, it is necessary to explain the application scenarios of the present invention. The present invention is a smart grid energy efficiency analysis and optimization system. In the actual operation of the smart grid, there is a high service communication delay due to unreasonable spectrum resource allocation, which increases the energy efficiency loss of the grid. By collecting smart grid operation status parameters and communication QoS demand parameters, and combining spectrum resource demand intensity coefficient analysis and optimization, it is ensured that spectrum resources are reasonably allocated, thereby ensuring that the spectrum resources of the smart grid are reasonably allocated, thereby reducing the energy efficiency loss of the smart grid.

[0056] Example 1: Figure 1 As shown, this embodiment provides a smart grid energy efficiency analysis and optimization system, characterized in that the system includes: a feature extraction module, a spectrum resource demand intensity coefficient sequence analysis module, a spectrum resource allocation module, and an energy efficiency optimization control module, and each module is sequentially connected in communication;

[0057] The feature extraction module is used to collect the smart grid operation status parameter set and the communication QoS requirement parameter set in real time through the sensor network deployed at the intersection of the physical layer and the communication layer of the smart grid, and extract the smart grid operation status feature vector after normalization through the sliding time window; the smart grid operation status parameter set includes voltage deviation and load mutation gradient, and the communication QoS requirement parameter set includes delay sensitivity, packet loss rate threshold, and service priority weight. The physical layer of the power grid includes substations and distribution cabinets, and the communication layer includes wireless base stations and optical fiber terminals. Use a sliding time window (such as a window every 5 seconds) to normalize the data, eliminate dimensional differences, and generate a standardized feature vector.

[0058] The spectrum resource demand intensity coefficient sequence analysis module is used to analyze the quantitative mapping relationship between the smart grid operation status and the communication QoS demand through the spatiotemporal association rule base of the smart grid operation status feature vector to obtain the benchmark spectrum resource demand intensity coefficient; the benchmark spectrum resource demand intensity coefficient is weighted according to the service priority weight to obtain the spectrum resource demand intensity coefficient sequence;

[0059] A spectrum resource allocation module is used to execute a spectrum resource allocation strategy according to the spectrum resource demand intensity coefficient sequence; the spectrum resource allocation strategy includes allocating a hard slice layer to protection and control services and a soft slice layer to metering and collection services; the spectrum resource allocation strategy is aligned with the timestamp of the smart grid control instruction through the time-sensitive network protocol, and it is ensured that the spectrum resource allocation is preloaded within a preset delay threshold before the control instruction is issued; the spectrum allocation strategy is aligned with the timestamp of the power grid control instruction through the time-sensitive network (TSN) protocol to ensure that the preloading is completed before the instruction is issued, such as 2ms in advance.

[0060] The energy efficiency optimization control module is used to obtain the communication delay of all services and calculate the delay standard deviation after the spectrum resource allocation strategy is executed, and construct the communication delay-energy efficiency loss mapping function by data analysis based on the delay standard deviation and energy efficiency index; optimize the spectrum resource allocation through data analysis to minimize the communication delay-energy efficiency loss mapping function, and the delay standard deviation is less than the preset safety delay standard deviation threshold. The communication delay-energy efficiency loss mapping function is used to quantify the impact of delay fluctuations on energy efficiency. The delay standard deviation is a performance characterization indicator of the spectrum resource allocation strategy.

[0061] The spectrum resource demand intensity coefficient sequence analysis module includes:

[0062] The coupling feature vector generation module is used to construct the communication link-power equipment association matrix based on the physical connection relationship between power equipment and communication links in the smart grid topology. ,express Power equipment and The connection relationship matrix between the communication links; among them, when the control instruction of the ath power equipment depends on the When a communication link is transmitted, ,otherwise ; , .

[0063] The incidence matrix Perform weighted processing to obtain the weighted correlation matrix ; The weight value of the weighted processing ,in, For the The electrical centrality of an electrical device, For the The number of transmission hops of a communication link.

[0064] The weighted incidence matrix and the smart grid operation status feature vector Perform tensor fusion to generate coupled feature vectors ; The smart grid operation status feature vector ,in, is the power imbalance, is the time window for load mutation, is the load mutation gradient; is the voltage deviation; is the delay sensitivity; is the packet loss rate threshold; For Business The priority weight of .

[0065] Spectrum resource demand intensity coefficient sequence calculation submodule, used to obtain the spatiotemporal association rule base },in, for The coupled eigenvector at time instant; = for Communication quality requirements at all times, including measured latency , Packet loss rate and spectrum resource demand intensity coefficient benchmark value .

[0066] The spectrum resource demand intensity coefficient benchmark value is obtained by training the spatiotemporal correlation model through the long short-term memory network. As the target value of the spatiotemporal correlation model, the network input is , output the predicted spectrum resource demand intensity coefficient , the loss function is: ,in, is the transient stability margin of the smart grid, is the preset weight coefficient;

[0067] The coupled eigenvector Input the trained spatiotemporal correlation model and output the spectrum resource demand intensity coefficient .

[0068] Based on business priority weight Spectrum resource demand intensity coefficient Make corrections to generate spectrum resource demand intensity coefficient sequences for each service ,in, , For Business The spectrum resource demand intensity coefficient is It is the sum of the priority weights of all services.

[0069] The association matrix represents the control dependency between power equipment and communication links. Assume that a smart grid contains 3 power equipment (l=3) and 2 communication links (k=2): the control instructions of equipment 1 depend on the transmission of communication link 1; the control instructions of equipment 2 depend on the transmission of communication link 2; the control instructions of equipment 3 do not depend on any communication link. Then the association matrix is: , assuming that the electrical centrality of device 1 is , the number of hops of communication link 1 , then the weight: , if the electrical center of device 2 is , the number of hops of communication link 2 , then the weight , the resulting weighted association matrix Grid transient stability margin constraints ≥10%. Electrical centering Through the calculation of node betweenness centrality of power grid topology, ,in For Node arrive The number of shortest paths, Passing Node The number of shortest paths. is the number of routing hops from the source node to the destination node of the data packet. The coupled eigenvector ensures that the spectrum resource allocation strategy responds to the physical state of the power grid at the same time ( ) and communication link characteristics ( ). Business priority weight It is calculated dynamically based on the historical fault frequency and real-time load rate of the smart grid, and is classified by business importance. Allocate spectrum resources to ensure that key control instructions of smart grids are given priority.

[0070] The spectrum resource requirement intensity coefficient sequence calculation submodule comprises the following steps:

[0071] The spectrum resource demand intensity benchmark value is obtained by using the spectrum resource demand intensity coefficient benchmark value calculation formula ; Spectrum resource demand intensity coefficient benchmark value The calculation formula is:

[0072] , and The rate of change is less than the preset Rate of change threshold.

[0073] in, , is the adaptive correction factor, , The initial value of is preset according to the smart grid topology; is the disturbance level factor of smart grid; is the response speed of the smart grid to disturbances, is the transient response time constant of the smart grid, The current timestamp.

[0074] The smart grid disturbance level factor The calculation formula is:

[0075] .

[0076] in, is the inertia time constant of the smart grid; is the power imbalance.

[0077] The adaptive correction factor makes adaptive adjustments to the load mutation gradient and voltage deviation respectively. It is pre-set according to the smart grid topology and adjusted according to the actual operation status of the grid to ensure that the allocation of spectrum resources meets the actual needs of the grid. k represents the response speed of the grid to disturbances; ΔP represents the power fluctuation in the grid, indicating the difference between supply and demand in the grid. The load mutation gradient represents the rate of change of grid power within a certain time window. The larger the value, the more drastic the load change of the grid. Therefore, more spectrum resources may be required for timely adjustment and communication. The voltage deviation represents the degree of deviation of the grid voltage. The voltage fluctuation will directly affect the stability of the grid. The larger the voltage deviation, the more spectrum resources are required to maintain the communication quality. The transient response time constant of the grid is calibrated through simulation or actual measurement. The smaller the smaller, the faster the grid response speed, and vice versa. When the smart grid is disturbed (such as power imbalance), the demand for spectrum resources needs to be adjusted according to the intensity of the disturbance. The disturbance level factor is calculated by the Sigmoid function. The Sigmoid function can dynamically adjust the spectrum demand according to the change of the power imbalance. The inertia time constant of the smart grid indicates the degree of slow response of the grid to disturbances. The larger the inertia time constant, the slower the grid responds to disturbances.

[0078] The spectrum resource allocation module includes:

[0079] The slice bandwidth management module is used for allocating the hard slice layer to the protection control service to allocate exclusive spectrum slices to the protection control service, and the hard slice bandwidth of the exclusive spectrum slice ,in, To protect bandwidth and meet ; is the total available spectrum bandwidth; For Business Hard slice bandwidth; is the sum of the spectrum resource demand intensity coefficients of all services;

[0080] The allocation of soft slice layers to metering and collection services is to use a non-orthogonal multiple access method for metering and collection services, and the soft slice bandwidth of the non-orthogonal multiple access method is ;in, is the dynamic interference suppression factor, .

[0081] The hard slice bandwidth and soft slice bandwidth constraints satisfy ,and in, The preset minimum protection bandwidth.

[0082] Protection and control services refer to some communication services that are critical to the stability and reliability of the power grid, such as real-time transmission of power grid control instructions, fault detection and emergency response. For these high-priority tasks, the system will allocate exclusive spectrum slices (hard slice layer) to ensure that these tasks have independent and sufficient spectrum bandwidth, thereby avoiding delays or packet loss caused by shared bandwidth. Exclusive spectrum slices provide zero-conflict channels for protection and control services. The delay fluctuation mainly comes from physical layer noise, and the delay standard deviation is low. The protection bandwidth is an additional bandwidth allocated to ensure that the bandwidth of protection and control services has sufficient redundancy. Metering and collection services usually refer to some tasks with low latency requirements but large-scale data transmission, such as smart meter data backhaul and load monitoring, and the demand for spectrum resources is relatively low. The non-orthogonal multiple access (NOMA) method is adopted, which is a more efficient access method that allows multiple devices to share bandwidth on the same spectrum to reduce interference. Non-orthogonal multiple access (NOMA) introduces co-channel interference, and the delay fluctuation is affected by power domain competition, and the delay standard deviation is high. When the hard slice bandwidth increases, the standard deviation of delay decreases, and the energy efficiency loss is reduced; when the soft slice interference suppression factor is optimized, the co-channel interference is reduced, and the standard deviation of delay is reduced. At the same time, the spectrum resource allocation strategy meets the energy efficiency constraint condition, which is that the slice bandwidth allocation of the exclusive spectrum slice is greater than the preset smart grid transient stability margin, and the dynamic interference suppression factor of the non-orthogonal multiple access method is greater than the preset smart grid transient stability margin. , To increase the renewable energy consumption rate; to ensure the renewable energy consumption rate When the bandwidth is increased, the interference suppression capability is enhanced and the bandwidth efficiency of soft slices is improved. The preset minimum protection bandwidth ensures that even under heavy load, protection control services can obtain sufficient bandwidth to ensure their real-time and reliability.

[0083] The spectrum resource allocation module further comprises the following steps:

[0084] The spectrum resource allocation strategy is aligned with the timestamp of the smart grid control instruction through the time-sensitive network protocol and a synchronization error is calculated, and the synchronization error is less than a preset synchronization error threshold.

[0085] Hard slice bandwidth preloading time window in spectrum resource allocation Transient response time constant of smart grid Satisfy constraints .

[0086] The control instructions in smart grids are usually triggered based on timestamps. In the process of smart grid control, certain errors may occur due to network delays, calculation time, etc., which are called synchronization errors. In order to ensure the accurate operation of the system, this error needs to be monitored and limited within a preset threshold to ensure that the allocation of spectrum resources is aligned with the time of the control instructions, so as to avoid the failure of the control instructions due to excessive delays. In the smart grid system, the control instructions need to be pre-configured and loaded before the spectrum resources are allocated to ensure that the control instructions can be executed quickly and accurately when they are issued. The pre-loading time window refers to the time period for pre-loading spectrum resources before the execution of the control instructions, ensuring that the spectrum pre-configuration takes effect before the completion of the transient process of the smart grid, so that the control instructions in the network are transmitted at the right time, thereby improving the real-time performance and scheduling efficiency of the system. When encountering disturbances (such as load changes, faults, etc.), the smart grid needs a certain amount of time to respond and stabilize. τ represents the time constant of this transient response, which measures the dynamic response speed of the power grid.

[0087] The energy efficiency optimization control module includes:

[0088] Energy efficiency analysis module, used for energy efficiency indicators including the reduction of grid backup capacity redundancy , Increase in renewable energy consumption rate ; The communication delay-energy efficiency loss mapping function Used to quantify the energy efficiency loss of the power grid according to the fluctuation of communication delay; communication delay-energy efficiency loss mapping function The calculation formula is:

[0089] .

[0090] in, is the standard deviation of the delay; is the normalized weight coefficient and satisfies .

[0091] Mapping function of communication delay-energy efficiency loss Input the reward function and dynamically adjust the spectrum allocation strategy to minimize it through the double-delayed deep deterministic policy gradient algorithm, and update the adaptive correction factor , .

[0092] By delay standard deviation The effect of communication delay fluctuation on power grid energy efficiency is quantified by the mapping function, which is a key indicator for optimizing power grid energy efficiency. In system optimization, the goal is to minimize the mapping function, thereby reducing the energy efficiency loss caused by communication delay fluctuation. , thereby reducing energy efficiency losses .pass calculate, is the baseline reserve capacity, is the optimized value; , is the benchmark value of renewable energy consumption rate, is the optimized value. The communication delay-energy efficiency loss mapping function is input into the reward function, with the purpose of dynamically adjusting the spectrum allocation strategy in the smart grid through the reward function to minimize the energy efficiency loss. The double-delay deep deterministic policy gradient (DDPG) algorithm is a reinforcement learning algorithm specifically used for continuous control problems. In the spectrum resource allocation of the smart grid, the system faces a continuous decision space. Therefore, the DDPG algorithm can continuously adjust the spectrum resource allocation strategy through interaction with the environment to minimize the reward function (i.e., reduce delay fluctuations and energy efficiency losses). Based on the feedback of the reward function, the DDPG algorithm will adjust the spectrum resource allocation strategy to optimize the energy efficiency performance of the power grid under different communication conditions. During the optimization process, the algorithm will also dynamically update the adaptive correction factor to adapt to different power grid conditions. The adjustment of the adaptive correction factor can affect the calculation of the spectrum resource demand intensity coefficient, thereby further optimizing the spectrum allocation strategy.

[0093] The energy efficiency analysis module comprises the following steps:

[0094] The reward function for:

[0095] .

[0096] in, Rewards for meeting the delay target; Penalty for exceeding the delay limit; Penalties for energy efficiency losses; reward for interference suppression; is the indicator function, when 1 if yes, 0 otherwise. is the preset safety delay standard deviation threshold; is a trainable hyperparameter that satisfies .

[0097] The value of is 1, indicating that the delay meets the standard and a reward is given; otherwise, it is 0, indicating that the delay does not meet the standard and no reward is given. Determines the intensity of the reward when the latency reaches the target. It is usually a large value to encourage the system to optimize latency to avoid the negative impact of communication latency on energy efficiency. For example, the initial value of the hyperparameter , and is optimized by gradient descent. The delay penalty ensures that the penalty only takes effect when the delay exceeds the limit. The interference suppression reward indicates that the better the interference suppression effect, the higher the reward. This constraint ensures the importance of the delay target reward in the total reward, because delay is crucial to the stability of the power grid, so the delay target reward occupies the main weight.

[0098] The energy efficiency analysis module also includes:

[0099] Adaptive correction factor update analysis module for adaptive correction factor Updates meet , adaptive correction factor Updates meet , ensuring the spectrum resource demand intensity benchmark value Related to the energy efficiency index. The update formula of the adaptive correction factor indicates that for every 10% increase in the renewable energy consumption rate, α increases by 1 time; for every 15% decrease in the reserve capacity redundancy, β decreases by 1 time.

[0100] The slice bandwidth management module also includes:

[0101] The energy efficiency priority arbitration module is used to trigger energy efficiency priority arbitration when the hard slice bandwidth demand exceeds the total spectrum resources.

[0102] The energy efficiency priority arbitration is to obtain the energy efficiency impact factor of each business , according to the energy efficiency influencing factor Allocate hard slice bandwidth in descending order until ∑ ≤ ; The energy efficiency influencing factor The calculation formula is: ;in, For Business The core purpose of energy efficiency priority arbitration is to sort the businesses according to the energy efficiency impact factor of each business, and give priority to those businesses that contribute more to the energy efficiency of the power grid to reduce the overall energy efficiency loss. For Business Renewable energy consumption contribution, an indicator to measure how the business helps smart grids absorb and utilize renewable energy.

[0103] Finally, it should be noted that: Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A smart grid energy efficiency analysis and optimization system, characterized in that: The system includes: a feature extraction module, a spectrum resource demand intensity coefficient sequence analysis module, a spectrum resource allocation module, and an energy efficiency optimization control module, and each module is sequentially connected in communication; The feature extraction module is used to collect the smart grid operation status parameter set and the communication QoS requirement parameter set in real time through the sensor network deployed at the intersection node of the smart grid physical layer and the communication layer, and extract the smart grid operation status feature vector after normalization through the sliding time window; The spectrum resource demand intensity coefficient sequence analysis module is used to analyze the quantitative mapping relationship between the smart grid operation status and the communication QoS demand through the spatiotemporal association rule base of the smart grid operation status feature vector to obtain the benchmark spectrum resource demand intensity coefficient; the benchmark spectrum resource demand intensity coefficient is weighted according to the service priority weight to obtain the spectrum resource demand intensity coefficient sequence; The spectrum resource demand intensity coefficient sequence analysis module includes: A coupling feature vector generation module is used to construct a communication link-power device association matrix M based on the physical connection relationship between the power device and the communication link in the smart grid topology; The association matrix M is weighted to obtain a weighted association matrix M'; The weighted correlation matrix M' is tensor-fused with the smart grid operation status feature vector F to generate the coupling feature vector The smart grid operation status feature vector Among them, ΔP is the power imbalance, Δt is the time window of load mutation, ΔP / Δt is the load mutation gradient; ΔV is the voltage deviation; τ max is the delay sensitivity; L max is the packet loss rate threshold; is the priority weight of service i, i∈[1,n]; The spectrum resource demand intensity coefficient sequence calculation submodule is used to obtain the spatiotemporal association rule base R = {(F t ',QoS t )|t=1,2,...,T}, where F t ' is the coupling characteristic vector at time t; QoS t =[τ act ,L act ,η r ] t is the communication quality requirement at time t, including the measured delay τ act , Packet loss rate L act and the spectrum resource demand intensity coefficient benchmark value η r ; The coupling feature vector F' is input into the trained spatiotemporal correlation model, and the spectrum resource demand intensity coefficient η is output; A spectrum resource allocation module, configured to execute a spectrum resource allocation strategy according to the spectrum resource demand intensity coefficient sequence; the spectrum resource allocation strategy includes allocating a hard slice layer to protection and control services and allocating a soft slice layer to metering and collection services; aligning the spectrum resource allocation strategy with the timestamp of the smart grid control instruction through a time-sensitive network protocol, and ensuring that the spectrum resource allocation is preloaded within a preset delay threshold before the control instruction is issued; Energy efficiency optimization control module, used to construct communication delay-energy efficiency loss mapping function by data analysis based on delay standard deviation and energy efficiency index; optimize spectrum resource allocation through data analysis to minimize communication delay-energy efficiency loss mapping function; The energy efficiency optimization control module includes: Energy efficiency analysis module, used for the energy efficiency indicators including the decrease in grid standby capacity redundancy ΔC and the increase in renewable energy consumption rate ΔE; the communication delay-energy efficiency loss mapping function G(σ τ ) is used to quantify the energy efficiency loss of the power grid according to the fluctuation of communication delay; the communication delay-energy efficiency loss mapping function G(σ τ ) is calculated as: Among them, σ τ is the delay standard deviation; c1, c2, c3 are normalized weight coefficients, and satisfy c1+c2+c3=1.

2. A smart grid energy efficiency analysis and optimization system according to claim 1, characterized in that: The spectrum resource demand intensity coefficient sequence analysis module includes: The correlation matrix M = [m ab ] l×k , represents the connection relationship matrix between l power devices and k communication links; when the control instruction of the ath power device depends on the bth communication link transmission, m ab =1, otherwise m ab =0; a∈[1,l], b∈[1,k]; The weight value of the weighted processing Among them, d a is the electrical centrality of the ath electrical equipment, h b is the transmission hop number of the bth communication link; The spectrum resource demand intensity coefficient benchmark value η is trained by long short-term memory network to obtain the spatiotemporal correlation model. r As the target value of the spatiotemporal correlation model, the network input is F t ', output the predicted spectrum resource demand intensity coefficient η', the loss function is: L S =||η'-η r ||2+λ·max(0,δ-10%), where δ is the transient stability margin of the smart grid and λ is the preset weight coefficient; Based on business priority weight The spectrum resource demand intensity coefficient η is modified to generate a spectrum resource demand intensity coefficient sequence {η1,...,η i ,...,η n },in, η i is the spectrum resource demand intensity coefficient of service i, It is the sum of the priority weights of all services.

3. A smart grid energy efficiency analysis and optimization system according to claim 2, characterized in that: The spectrum resource requirement intensity coefficient sequence calculation submodule comprises the following steps: The spectrum resource demand intensity benchmark value η is obtained by using the spectrum resource demand intensity coefficient benchmark value calculation formula r ; Spectrum resource demand intensity coefficient benchmark value η r The calculation formula is: η r =α·(ΔP / Δt)·γ+β·|ΔV|·e -t / τ , 0<η r ≤1 and η r The rate of change is less than the preset η r rate of change threshold; Among them, α and β are adaptive correction factors, and the initial values ​​of α and β are preset according to the smart grid topology structure; γ is the smart grid disturbance level factor; e -t / τ is the response speed of the smart grid to disturbances, τ is the transient response time constant of the smart grid, and t is the current timestamp; The calculation formula of the smart grid disturbance level factor γ is: Where k is the inertia time constant of the smart grid; ΔP is the power imbalance.

4. The smart grid energy efficiency analysis and optimization system according to claim 2, characterized in that: The spectrum resource allocation module includes: The slice bandwidth management module is used for allocating the hard slice layer to the protection control service to allocate exclusive spectrum slices to the protection control service, and the hard slice bandwidth of the exclusive spectrum slice Where ΔB is the protection bandwidth and satisfies B t is the total available spectrum bandwidth; is the hard slice bandwidth of service i; is the sum of the spectrum resource demand intensity coefficients of all services; The allocation of soft slice layers to metering and collection services is to use a non-orthogonal multiple access method for metering and collection services, and the soft slice bandwidth of the non-orthogonal multiple access method is Where ξ is the dynamic interference suppression factor, 0<ξ<1; The hard slice bandwidth and soft slice bandwidth constraints satisfy and Among them, B min The preset minimum protection bandwidth.

5. A smart grid energy efficiency analysis and optimization system according to claim 4, characterized in that: The spectrum resource allocation module further comprises the following steps: The spectrum resource allocation strategy is aligned with the timestamp of the smart grid control instruction through the time-sensitive network protocol and a synchronization error is calculated, wherein the synchronization error is less than a preset synchronization error threshold; Hard slice bandwidth preloading time window t in spectrum resource allocation pre The transient response time constant τ of the smart grid satisfies the constraint condition t pre ≤τ / 5.

6. The smart grid energy efficiency analysis and optimization system according to claim 1, characterized in that: The energy efficiency optimization control module includes: The communication delay-energy efficiency loss mapping function G(σ τ ) inputs the reward function and dynamically adjusts the spectrum allocation strategy to minimize it through the double-delayed deep deterministic policy gradient algorithm, and updates the adaptive correction factors α and β.

7. A smart grid energy efficiency analysis and optimization system according to claim 6, characterized in that: The energy efficiency analysis module comprises the following steps: The reward function R is: in, is the delay reaching reward; ρ1·max(0,σ τ -σ th ) is the penalty for delay exceeding limit; ρ2·G(σ τ ) is the energy efficiency loss penalty; ρ3·log(1-ξ) is the interference suppression reward; is the indicator function, when σ τ ≤σ th 1 when σ is equal to 1, otherwise 0. th is the preset safety delay standard deviation threshold; μ, ρ1, ρ2, ρ3 are trainable hyperparameters that satisfy μ>>ρ1+ρ2.

8. The smart grid energy efficiency analysis and optimization system according to claim 6, characterized in that: The energy efficiency analysis module also includes: Adaptive correction factor update analysis module, used to update the adaptive correction factor α to satisfy α t+1 =α t (1+ΔE / 10%), the update of the adaptive correction factor β satisfies β t+1 =β t (1-ΔC / 15%), to ensure the spectrum resource demand intensity benchmark value η r Related to energy efficiency indicators.

9. The smart grid energy efficiency analysis and optimization system according to claim 4, characterized in that: The slice bandwidth management module also includes: Energy efficiency priority arbitration module, used to trigger energy efficiency priority arbitration when the hard slice bandwidth demand exceeds the total spectrum resources; The energy efficiency priority arbitration is to obtain the energy efficiency impact factor ∈ of each service i , according to the energy efficiency impact factor ∈ i Allocate hard slice bandwidth in descending order until the The energy efficiency impact factor ∈ i The calculation formula is: ∈ i =η i ΔE i ; where ΔE i Contribute to the renewable energy consumption of business i.

Images