Control method and device for participating in power grid demand response based on 5g base station resources

Abstract

The application discloses a control method and device for participating in power grid demand response based on 5G base station resources, comprising the following steps: each base station and a power-using family are connected to a power distribution network; a target function is constructed based on the requirement of minimizing system power fluctuation; a base station transmission data prediction model is constructed; the target function is converted and a base station communication service quality constraint condition is constructed; a deep LSTM load prediction model is constructed, and part of unknown parameters in the target function are solved; an unexpected constraint is constructed, and the target function is adjusted to a target function containing the unexpected constraint; decision variables that satisfy the base station communication service quality constraint condition and make the target function containing the unexpected constraint minimum are calculated, a predicted next-time data packet size state to be transmitted is obtained, and the base station transmits data packets to users based on the prediction result; the application has the advantages that the demand of minimizing power adjustment is met, and thus the power grid demand response effect is good.

Description

Technical Field

[0001] This invention relates to the fields of 5G downlink communication data transmission and power grid demand response, specifically to a control method and apparatus for participating in power grid demand response based on 5G base station resources. Background Technology

[0002] Based on its high bandwidth, wide connectivity, and low latency, 5G mobile communication technology offers superior communication quality and richer application scenarios. The ultra-high frequency of 5G will lead to shorter transmission distances and higher base station density. This results in a rapid increase in the power consumption of 5G base stations, approximately three times that of 4G base stations. As the main energy-consuming component of cellular wireless networks, 5G base stations may account for about 70% of the total energy consumption of the entire communication network equipment. Demand response is a crucial aspect of electricity demand-side management. Due to the sheer number and flexibility of 5G base stations, they represent a considerable potential resource for demand response. Based on the signal-to-noise ratio (SNR) and the modulation and coding scheme selected accordingly, base stations can allocate resource blocks with different schemes to the data packets to be transmitted, thereby adjusting the corresponding transmit power. Therefore, reasonable optimization and control of the 5G network can be implemented to regulate the downlink power consumption of 5G base stations to participate in the demand response optimization operation of the active distribution network.

[0003] Chinese Patent Publication No. CN114977163A discloses a wireless resource allocation method for active distribution network demand response based on 5G networks, including: 1. constructing an active distribution network downlink network environment; 2. encoding transmitted data packets according to base station number, user number, and data packet type; 3. constructing a mixed-integer linear programming model with linear constraints, using minimum power consumption fluctuation as the objective function; 4. establishing a probabilistic constrained programming model under different user demand categories using a sample average approximation strategy; 5. predicting user household power load using a long short-term memory neural network; 6. constructing a globally optimal allocation scheme by combining the Lagrange dual algorithm and the trust region algorithm, allocating an optimal modulation and coding scheme and resource block to each data packet. This patent application can achieve peak shaving and valley filling of the active distribution network by 5G base stations while ensuring user service quality, and effectively reduce the system peak-valley difference and packet loss rate. However, this patent application cannot meet the requirement of minimizing power regulation, resulting in insufficient grid demand response performance. Summary of the Invention

[0004] The technical problem to be solved by the present invention is that the existing power grid demand response optimization control method cannot meet the requirement of minimizing power regulation, thus the power grid demand response effect is not good enough.

[0005] This invention solves the above-mentioned technical problems through the following technical means: a control method based on 5G base station resources participating in power grid demand response, comprising the following steps:

[0006] Step 1: Connect each base station and household to the power distribution network and measure power consumption;

[0007] Step 2: Construct the objective function based on the requirement of minimizing system power fluctuations;

[0008] Step 3: Based on the size of the data packet to be transmitted at the current moment, predict the size of the data packet to be transmitted at the next moment, and construct a base station data transmission prediction model;

[0009] Step 4: Transform the objective function and construct the base station communication service quality constraints;

[0010] Step 5: Construct a deep LSTM load forecasting model and solve for some unknown parameters in the objective function;

[0011] Step 6: Construct unexpected constraints and adjust the objective function to include the unexpected constraints;

[0012] Step 7: Calculate the decision variable that satisfies the base station's communication service quality constraints and minimizes the objective function containing unexpected constraints. Substitute this decision variable into the base station's data transmission prediction model to obtain the predicted size of the data packet to be transmitted at the next moment. The base station then transmits the data packet to the user based on this prediction result.

[0013] Furthermore, the characteristic feature is that step one includes:

[0014] Each base station and household is connected to the power distribution network, and smart meters are used to measure power consumption. The base stations provide downlink communication services to users. The total power consumption of the power distribution network consists of the downlink communication power consumption provided by the base stations and the users' power load. Assume there are B 5G base stations in the system, each serving N users. Considering K types of user traffic and M types of MCS selection methods, define the set of demand response control cycles as T, where t represents the t-th control cycle in T, and the duration of the control cycle is DT: Let D... b，n，k，t Y represents the size of the k-th type of traffic data packet transmitted by user n within base station b during the t-th control period; Y is the total number of resource blocks (RBs) for each time slot.

[0015] Furthermore, step two includes:

[0016] The objective function is constructed using the following formula.

[0017]

[0018]

[0019]

[0020] in, These represent the current power of the base station and the household during the control period, and the predicted power of the base station and the household during the entire intraday control period, respectively; x b，n，k，m，t It is a decision variable in binary form, representing the selection of the m-th MCS to send data packets when the b-th base station provides downlink transmission service to the n-th user in the t-th control period. This represents the power consumed and transmission speed on each RB when sending data packets in the m-th MCS mode.

[0021] Furthermore, step three includes:

[0022] The current size of the data packet to be transmitted is as follows:

[0023]

[0024] Through formula

[0025]

[0026] Predict the size of the data packets to be transmitted in the next moment and construct a base station data transmission prediction model;

[0027] in, This represents the size of the new data packet that needs to be transmitted at time t.

[0028] Furthermore, step four includes:

[0029] Through formula The objective function is transformed, where,

[0030]

[0031] Construct the base station communication service quality constraints according to the following formula.

[0032]

[0033]

[0034]

[0035]

[0036]

[0037] x b，n，k，m，t ∈{0,1}

[0038] Where α represents the path loss compensation factor; PL b，nThis refers to the downlink path loss measured at a 5G base station; μ represents thermal interference noise; SINR m It is the signal-to-noise ratio of the m-th MCS selection; P max This refers to the maximum dynamic power of the base station. Minimum transmission speed for each type of traffic communication.

[0039] Furthermore, step five includes:

[0040] The deep LSTM load prediction model comprises a sequentially connected input layer, multiple hidden layers, and an output layer, which calculates the current power P of base stations and homes during the control cycle. t S P t H Inputting a deep LSTM load forecasting model yields the predicted power P of base stations and homes throughout the entire intraday control cycle. t PS P t PH .

[0041] Furthermore, step six includes:

[0042] Through formula Construct unexpected constraints;

[0043] Through formula The objective function is adjusted to include unexpected constraints, where I represents the total number of samples to be regulated in each regulation cycle. This represents the i-th decision component in sample I within the regulation period.

[0044] This invention also provides a control device for participating in power grid demand response based on 5G base station resources, comprising:

[0045] The initial setup module is used to connect each base station and household to the power distribution network and measure power consumption.

[0046] The objective function construction module is used to construct an objective function based on the requirement of minimizing system power fluctuations.

[0047] The prediction model building module is used to predict the size of the data packet to be transmitted at the next moment based on the size of the data packet to be transmitted at the current moment, and to build a prediction model for base station data transmission.

[0048] The constraint setting module is used to transform the objective function and construct the base station communication service quality constraints.

[0049] The parameter solving module is used to construct a deep LSTM load forecasting model and solve for some unknown parameters in the objective function.

[0050] The objective function adjustment module is used to construct unexpected constraints and adjust the objective function to include the unexpected constraints.

[0051] The prediction result output module is used to calculate the decision variable that satisfies the base station communication service quality constraints and minimizes the objective function containing unexpected constraints. The decision variable is substituted into the base station transmission data prediction model to obtain the predicted size of the data packet to be transmitted at the next moment. The base station transmits data packets to users based on the prediction result.

[0052] Furthermore, the initial setup module is also used for:

[0053] Each base station and household electricity user is connected to the power distribution network, and smart meters are used to measure power consumption. The base stations provide downlink communication services to users. The total power consumption of the power distribution network consists of the downlink communication power consumption provided by the base stations and the users' power load. Assume there are B 5G base stations in the system, each serving N users. Considering K types of user traffic and M types of MCS selection methods, define the set of demand response control cycles as T, where t represents the t-th control cycle in T, and the duration of the control cycle is DT. Let D... b，n，k，t Y represents the size of the k-th type of traffic data packet transmitted by user n within base station b during the t-th control period; Y is the total number of resource blocks (RBs) for each time slot.

[0054] Furthermore, the objective function construction module is also used for:

[0055] The objective function is constructed using the following formula.

[0056]

[0057]

[0058]

[0059] in, These represent the current power of the base station and the household during the control period, and the predicted power of the base station and the household during the entire intraday control period, respectively; x b，n，k，m，t It is a decision variable in binary form, representing the selection of the m-th MCS to send data packets when the b-th base station provides downlink transmission service to the n-th user in the t-th control period. R m This represents the power consumed and transmission speed on each RB when sending data packets in the m-th MCS mode.

[0060] Furthermore, the prediction model building module is also used for:

[0061] The current size of the data packet to be transmitted is as follows:

[0062]

[0063] Through formula

[0064]

[0065] Predict the size of the data packets to be transmitted in the next moment and construct a base station data transmission prediction model;

[0066] in, This represents the size of the new data packet that needs to be transmitted at time t.

[0067] Furthermore, the constraint setting module is also used for:

[0068] Through formula The objective function is transformed, where,

[0069]

[0070] Construct the base station communication service quality constraints according to the following formula.

[0071]

[0072]

[0073]

[0074]

[0075]

[0076] x b，n，k，m，t ∈{0,1}

[0077] Where α represents the path loss compensation factor; PL b,n This refers to the downlink path loss measured at a 5G base station; μ represents thermal interference noise; SINR m It is the signal-to-noise ratio of the m-th MCS selection; P max This refers to the maximum dynamic power of the base station. Minimum transmission speed for each type of traffic communication.

[0078] Furthermore, the parameter solving module is also used for:

[0079] The deep LSTM load prediction model comprises a sequentially connected input layer, multiple hidden layers, and an output layer, which calculates the current power P of base stations and homes during the control cycle. t S P tH Inputting a deep LSTM load forecasting model yields the predicted power P of base stations and homes throughout the entire intraday control cycle. t PS P t PH .

[0080] Furthermore, the objective function adjustment module is also used for:

[0081] Through formula Construct unexpected constraints;

[0082] Through formula The objective function is adjusted to include unexpected constraints, where I represents the total number of samples to be regulated in each regulation cycle. This represents the i-th decision component in sample I within the regulation period.

[0083] The advantages of this invention are:

[0084] (1) The present invention constructs an objective function based on the requirement of minimizing system power fluctuations, and transforms and adjusts the objective function to obtain an objective function containing unexpected constraints. Finally, the decision variables that satisfy the base station communication service quality constraints and minimize the objective function containing unexpected constraints are solved. The objective function is set with other constraints on the premise of satisfying the requirement of minimizing power regulation, which further improves the grid demand response effect. Overall, the grid demand response effect is good.

[0085] (2) This invention considers utilizing the intangible communication resources of 5G base stations as a means of participating in power grid demand response, achieving optimized regulation of information and energy integration, and realizing cost reduction and efficiency improvement; it adopts a prediction-optimization-regulation rolling framework based on the optimized allocation of 5G downlink communication resources. This framework comprehensively considers constraints such as ultra-short-term household load forecasting, base station downlink power regulation, resource allocation, and user communication service quality. Through this rolling framework, household load can be continuously predicted, resource allocation optimized, and base stations can be regulated at the minute level.

[0086] (3) The literature on background technology cannot meet the needs of power grid minimizing power fluctuation regulation. It cannot make real-time dynamic adjustments based on the current daily electricity consumption data and communication load size to meet the rolling optimization regulation of the power grid in each regulation cycle and achieve the optimal demand response throughout the day. However, the present invention meets the needs of power grid minimizing power fluctuation regulation and makes real-time dynamic adjustments based on the current daily electricity consumption data and communication load size to meet the rolling optimization regulation of the power grid in each regulation cycle and achieve the optimal demand response throughout the day. Attached Figure Description

[0087] Figure 1This is a flowchart of a control method for participating in power grid demand response based on 5G base station resources, as disclosed in an embodiment of the present invention.

[0088] Figure 2 This is a schematic diagram of base station communication resource allocation according to power grid demand response requirements in the control method based on 5G base station resources participating in power grid demand response disclosed in the embodiments of the present invention. Detailed Implementation

[0089] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0090] Example 1

[0091] like Figure 1 and Figure 2 As shown, this invention provides a control method for power grid demand response based on 5G base station resources. It employs a multi-stage objective dynamic programming method based on dynamic programming principles, introducing the concept of phased rolling optimization. This method updates the load in real time to optimize the system model and rolls it over a full-day timescale for regulation, constructing a prediction-optimization-regulation rolling framework. The method includes the following steps:

[0092] S1. Each base station and household is connected to the power distribution network, and power consumption is measured; the specific process is as follows:

[0093] Each base station and household electricity user is connected to the power distribution network and uses smart meters to measure power consumption. Information is uploaded to the control platform via smart data acquisition terminals. Base stations provide downlink communication services to users, supporting their text, video, and voice needs. The total power consumption of the power distribution network consists of the downlink communication power consumption provided by the base stations and the users' power load. By rationally controlling the former, peak shaving and valley filling of the system's total power consumption can be achieved. Assuming there are B 5G base stations in the system, each serving N users, and considering K types of user traffic and M types of MCS selection methods, the set of demand response control cycles is defined as T, where t represents the t-th control cycle in T, and the duration of the control cycle is DT. Let D... b，n，k，t Y represents the size of the k-th type of traffic data packet transmitted by user n within base station b during the t-th control period; Y is the total number of resource blocks (RBs) for each time slot.

[0094] S2. Construct the objective function based on the requirement of minimizing system power fluctuations; the specific process is as follows:

[0095] Using equations (1)-(3), the optimization framework for 5G base stations participating in demand response is modeled as a problem of minimizing system power consumption fluctuations. The power of the base station will participate in the grid's demand response, where the fluctuation J of system power consumption is defined as the sum of the absolute differences between the sum of the base station downlink power consumption and the household electricity load in each control cycle and the average daily power consumption.

[0096]

[0097]

[0098]

[0099] In formula (1) This represents the current power of base stations and households during the control cycle and the predicted power for the entire intraday control cycle; in equation (3), x b，n，k，m，t It is a decision variable in binary form, representing the selection of the m-th MCS to send data packets when the b-th base station provides downlink transmission service to the n-th user in the t-th control period. R m This represents the power consumed and transmission speed on each RB when sending data packets in the m-th MCS mode.

[0100] S3. Based on the size of the data packet to be transmitted at the current moment, predict the size of the data packet to be transmitted at the next moment, and construct a base station data transmission prediction model; the specific process is as follows:

[0101] Within a control cycle, the size D(t) of the data packets transmitted by the base station to the user consists of the following two parts: the remaining data packet size DS after transmission at the previous moment due to the limitation of the number of channels Y. R And the newly added data packet size DS to be transmitted at this moment N This can be expressed as equations (4)-(6):

[0102]

[0103]

[0104]

[0105] Where D b，n，k，t-1 These are the components of the total data packet size D(t-1) of the previous control cycle.

[0106] Thus, we obtain the current state of the data packet size to be transmitted, as shown in equation (7):

[0107]

[0108] Therefore, the spatial state of the model predictive control is obtained from equations (8)-(10):

[0109]

[0110]

[0111] S4. Transform the objective function and construct the base station communication service quality constraints; the specific process is as follows:

[0112] The objective function is transformed into equations (10)-(11):

[0113]

[0114]

[0115] Where J is the overall objective function for minimizing the system, J t It is the optimization goal at each step; such as Figure 2 As shown, by allocating 5G base station communication resources in real time to regulate the downlink power consumption of the base station, changing the total power of the power grid, and thus participating in the power grid demand response, a prediction-optimization-regulation rolling framework is constructed.

[0116] The base station communication service quality is constrained according to formulas (12)-(16), and the user's communication service quality needs are met according to channel conditions and different types of traffic rate requirements.

[0117]

[0118]

[0119]

[0120]

[0121]

[0122] x b，n，k，m，t ∈{0,1} (16)

[0123] In equation (12), α represents the path loss compensation factor; PL b，n SINR is the downlink path loss measured by the 5G base station, which is related to the distance from the UE to the base station; μ represents thermal interference noise; m It is the signal-to-noise ratio of the m-th MCS selection; in equation (13) P max The maximum dynamic power of the base station; in equation (15) Minimum transmission speed for each type of traffic communication.

[0124] S5. Construct a deep LSTM load forecasting model and solve for some unknown parameters in the objective function; the specific process is as follows:

[0125] A regional-level ultra-short-term load forecasting method based on deep long short-term memory (LSTM) networks is adopted to predict household electricity load for the next 15 minutes to obtain various load values ​​required in the model. First, the input dataset is preprocessed and divided into training, validation, and test sets. A deep LSTM load forecasting model is composed of an input layer, multiple hidden layers, and an output layer. The input and hidden layers jointly extract features from the input data, and a fully connected layer is used as the output of the predicted household electricity load. Finally, a random search method is used to find suitable hyperparameters until the prediction error on the test set is minimized. In this embodiment, the current power P of the base station and households during the control cycle is used. t S P t H Inputting a deep LSTM load forecasting model yields the predicted power P of base stations and homes throughout the entire intraday control cycle. t PS P t PH .

[0126] S6. Construct unexpected constraints and adjust the objective function to include these unexpected constraints; the specific process is as follows:

[0127] The established prediction-optimization-regulation rolling framework is a mixed-integer linear programming model with linear constraints. Unexpected constraints are constructed through equation (17):

[0128]

[0129] Using equation (18), the objective function J t Represented as an objective function containing unexpected constraints:

[0130]

[0131] S7. Based on the lp-solve solver, the iterative algorithm for searching the globally optimal allocation scheme is used to solve the mixed-integer linear programming model, thereby obtaining the optimal data scheduling and resource allocation scheme for the base station that meets the power grid demand response requirements. Specifically, the decision variable that satisfies the base station communication service quality constraints and minimizes the objective function containing unexpected constraints is calculated. This decision variable is substituted into the base station transmission data prediction model to obtain the predicted size of the data packets to be transmitted at the next moment. The base station transmits data packets to users based on this prediction result. It should be noted that the solution process of the lp-solve solver is existing technology. The solution method in step six of the literature described in the background technology can be referred to. A Lagrange multiplier λ is introduced, and all unexpected and random constraints are put into the objective function with Lagrange multipliers through the Lagrange relaxation method. The search direction d of the iterative algorithm for searching the globally optimal allocation scheme is determined, and an iterative algorithm for searching the optimal Lagrange multiplier is constructed. A search step size based on the trust region method is adopted. If the difference between the objective values ​​of the last two iterations is less than a certain threshold, the step size is increased; otherwise, the step size is decreased. Furthermore, given that the model has been constructed and the objective function and constraints are clear, any other existing solution method can be used to solve the problem. Therefore, the solution method will not be elaborated here.

[0132] Through the above technical solutions, this invention constructs an objective function based on the requirement of minimizing system power fluctuations. The objective function is then transformed and adjusted to obtain an objective function containing unexpected constraints. Finally, the decision variables that satisfy the base station communication service quality constraints and minimize the objective function containing unexpected constraints are solved. The objective function, while meeting the requirement of minimizing power regulation, also sets other constraints to further improve the grid demand response effect. Overall, the grid demand response effect is good. Constraints on ultra-short-term household load forecasting, base station downlink power regulation, resource allocation, and user communication service quality are considered. Through this rolling framework, household load can be continuously predicted, resource allocation optimized, and base stations adjusted at the minute level. An iterative algorithm based on the Lagrange dual algorithm is used for the global optimal allocation scheme, realizing large-scale combinatorial optimization problems under complex constraints and reducing the computation time when solving the problem.

[0133] Example 2

[0134] Based on Embodiment 1, Embodiment 2 of the present invention also provides a control device for participating in power grid demand response based on 5G base station resources, including:

[0135] The initial setup module is used to connect each base station and household to the power distribution network and measure power consumption.

[0136] The objective function construction module is used to construct an objective function based on the requirement of minimizing system power fluctuations.

[0137] The prediction model building module is used to predict the size of the data packet to be transmitted at the next moment based on the size of the data packet to be transmitted at the current moment, and to build a prediction model for base station data transmission.

[0138] The constraint setting module is used to transform the objective function and construct the base station communication service quality constraints.

[0139] The parameter solving module is used to construct a deep LSTM load forecasting model and solve for some unknown parameters in the objective function.

[0140] The objective function adjustment module is used to construct unexpected constraints and adjust the objective function to include the unexpected constraints.

[0141] The prediction result output module is used to calculate the decision variable that satisfies the base station communication service quality constraints and minimizes the objective function containing unexpected constraints. The decision variable is substituted into the base station transmission data prediction model to obtain the predicted size of the data packet to be transmitted at the next moment. The base station transmits data packets to users based on the prediction result.

[0142] Specifically, the initial setup module is also used for:

[0143] Each base station and household electricity user is connected to the power distribution network, and smart meters are used to measure power consumption. The base stations provide downlink communication services to users. The total power consumption of the power distribution network consists of the downlink communication power consumption provided by the base stations and the users' power load. Assume there are B 5G base stations in the system, each serving N users. Considering K types of user traffic and M types of MCS selection methods, define the set of demand response control cycles as T, where t represents the t-th control cycle in T, and the duration of the control cycle is DT. Let D... b，n，k，t Y represents the size of the k-th type of traffic data packet transmitted by user n within base station b during the t-th control period; Y is the total number of resource blocks (RBs) for each time slot.

[0144] More specifically, the objective function construction module is also used for:

[0145] The objective function is constructed using the following formula.

[0146]

[0147]

[0148]

[0149] in, These represent the current power of the base station and the household during the control period, and the predicted power of the base station and the household during the entire intraday control period, respectively; xb，n，k，m，t It is a decision variable in binary form, representing the selection of the m-th MCS to send data packets when the b-th base station provides downlink transmission service to the n-th user in the t-th control period. R m This represents the power consumed and transmission speed on each RB when sending data packets in the m-th MCS mode.

[0150] More specifically, the prediction model building module is also used for:

[0151] The current size of the data packet to be transmitted is as follows:

[0152]

[0153] Through formula

[0154]

[0155] Predict the size of the data packets to be transmitted in the next moment and construct a base station data transmission prediction model;

[0156] in, This represents the size of the new data packet that needs to be transmitted at time t.

[0157] More specifically, the constraint setting module is also used for:

[0158] Through formula The objective function is transformed, where,

[0159]

[0160] Construct the base station communication service quality constraints according to the following formula.

[0161]

[0162]

[0163]

[0164]

[0165]

[0166] x b，n，k，m，t ∈{0,1}

[0167] Where α represents the path loss compensation factor; PL b，n This refers to the downlink path loss measured at a 5G base station; μ represents thermal interference noise; SINR m It is the signal-to-noise ratio of the m-th MCS selection; Pmax This refers to the maximum dynamic power of the base station. Minimum transmission speed for each type of traffic communication.

[0168] More specifically, the parameter solving module is also used for:

[0169] The deep LSTM load prediction model comprises a sequentially connected input layer, multiple hidden layers, and an output layer, which calculates the current power of base stations and homes during the control cycle. Input a deep LSTM load forecasting model to obtain the predicted power of base stations and homes throughout the entire intraday control cycle.

[0170] More specifically, the objective function adjustment module is also used for:

[0171] Through formula Construct unexpected constraints;

[0172] Through formula The objective function is adjusted to include unexpected constraints, where I represents the total number of samples to be regulated in each regulation cycle. This represents the i-th decision component in sample I within the regulation period.

[0173] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A control method based on 5G base station resources participating in power grid demand response, characterized in that, Includes the following steps: Step 1: Each base station and household with electricity is connected to the power distribution network, and power consumption is measured; Step 1 includes: Each base station and household electricity user is connected to the power distribution network, and smart meters are used to measure power consumption. The base stations provide downlink communication services to users. The total power consumption of the power distribution network consists of the downlink communication power consumption provided by the base stations and the users' power load. Assume there are B 5G base stations in the system, each serving N users, and consider K types of user traffic and M types of MCS selection methods. Define the set of demand response control cycles as T, where t represents the t-th control cycle in T, and the duration of the control cycle is DT. Let... This represents the size of the k-th type of traffic data packet transmitted by user n within base station b during the t-th control period; The total number of resource blocks (RBs) for each time slot; Step 2: Construct an objective function based on the requirement of minimizing system power fluctuations; Step 2 includes: The objective function is constructed using the following formula. in, These represent the current power of the base station and the household during the control period, and the predicted power of the base station and the household throughout the entire intraday control period, respectively. It is a decision variable in binary form, representing the selection of the m-th MCS to send data packets when the b-th base station provides downlink transmission service to the n-th user in the t-th control period. This represents the power consumed and transmission speed on each RB when sending data packets using the m-th MCS method; Step 3: Based on the current data packet size, predict the next data packet size to be transmitted, thus constructing a base station data transmission prediction model; Step 3 includes: The current size of the data packet to be transmitted is as follows: Through formula Predict the size of the data packets to be transmitted in the next moment and construct a base station data transmission prediction model; in, This represents the size of the new data packet that needs to be transmitted at time t; Step four: Transform the objective function and construct the base station communication service quality constraints; Step four includes: Through formula The objective function is transformed, where, Construct the base station communication service quality constraints according to the following formula. in, Indicates the path loss compensation factor; This is the downlink path loss measured by a 5G base station; Indicates thermal interference noise; It is the first Select the signal-to-noise ratio for each MCS; This refers to the maximum dynamic power of the base station. Minimum transmission speed for each type of traffic communication; Step 5: Construct a deep LSTM load forecasting model and solve for some unknown parameters in the objective function; Step 6: Construct unexpected constraints and adjust the objective function to include the unexpected constraints; Step 7: Calculate the decision variable that satisfies the base station communication service quality constraints and minimizes the objective function containing unexpected constraints. Substitute this decision variable into the base station transmission data prediction model to obtain the predicted size of the data packet to be transmitted at the next moment. The base station transmits data packets to the user based on the prediction results of the base station transmission data prediction model.

2. The control method based on 5G base station resources participating in power grid demand response according to claim 1, characterized in that, Step five includes: The deep LSTM load prediction model comprises a sequentially connected input layer, multiple hidden layers, and an output layer, which calculates the current power of base stations and homes during the control cycle. , Input a deep LSTM load forecasting model to obtain the predicted power of base stations and homes throughout the entire intraday control cycle. , .

3. The control method based on 5G base station resources participating in power grid demand response according to claim 2, characterized in that, Step six includes: Through formula Construct unexpected constraints; Through formula The objective function is adjusted to include unexpected constraints, where... This represents the total number of samples to be regulated in each regulation cycle. This represents the i-th decision component in sample I within the regulation period.

4. A control device for participating in power grid demand response based on 5G base station resources, characterized in that, include: The initial setup module is used to connect each base station and household to the power distribution network and measure power consumption; the initial setup module is also used for: Each base station and household electricity user is connected to the power distribution network, and smart meters are used to measure power consumption. The base stations provide downlink communication services to users. The total power consumption of the power distribution network consists of the downlink communication power consumption provided by the base stations and the users' power load. Assume there are B 5G base stations in the system, each serving N users, and consider K types of user traffic and M types of MCS selection methods. Define the set of demand response control cycles as T, where t represents the t-th control cycle in T, and the duration of the control cycle is DT. Let... This represents the size of the k-th type of traffic data packet transmitted by user n within base station b during the t-th control period; The total number of resource blocks (RBs) for each time slot; The objective function construction module is used to construct an objective function based on the requirement of minimizing system power fluctuations; the objective function construction module is also used for: The objective function is constructed using the following formula. in, These represent the current power of the base station and the household during the control period, and the predicted power of the base station and the household throughout the entire intraday control period, respectively. It is a decision variable in binary form, representing the selection of the m-th MCS to send data packets when the b-th base station provides downlink transmission service to the n-th user in the t-th control period. This represents the power consumed and transmission speed on each RB when sending data packets using the m-th MCS method; The prediction model building module is used to predict the size of the data packet to be transmitted at the next moment based on the size of the data packet to be transmitted at the current moment, thereby constructing a base station data transmission prediction model; the prediction model building module is also used for: The current size of the data packet to be transmitted is as follows: Through formula Predict the size of the data packets to be transmitted in the next moment and construct a base station data transmission prediction model; in, This represents the size of the new data packet that needs to be transmitted at time t; The constraint setting module is used to transform the objective function and construct the base station communication service quality constraints; the constraint setting module is also used for: Through formula The objective function is transformed, where, Construct the base station communication service quality constraints according to the following formula. in, Indicates the path loss compensation factor; This is the downlink path loss measured by a 5G base station; Indicates thermal interference noise; It is the first Select the signal-to-noise ratio for each MCS; This refers to the maximum dynamic power of the base station. Minimum transmission speed for each type of traffic communication; The parameter solving module is used to construct a deep LSTM load forecasting model and solve for some unknown parameters in the objective function. The objective function adjustment module is used to construct unexpected constraints and adjust the objective function to include the unexpected constraints. The prediction result output module is used to calculate the decision variable that satisfies the base station communication service quality constraints and minimizes the objective function containing unexpected constraints. The decision variable is substituted into the base station transmission data prediction model to obtain the predicted size of the data packet to be transmitted at the next moment. The base station transmits data packets to the user based on the prediction results of the base station transmission data prediction model.

Images