Intelligent remote power cut and power transmission control method based on Internet of Things

Abstract

The invention discloses an intelligent remote power cut and power transmission control method based on the Internet of Things. The method comprises the following steps: acquiring load change and power demand of each electric device of a power grid in real time by using the Internet of Things technology; establishing a nonlinear mapping relationship between the load change and the power demand by using a weighted optimization extreme learning machine, and outputting a prediction result of a future load demand; inputting the prediction result into a control model constructed based on a multiple-input multiple-output fuzzy control algorithm to obtain a power failure and power transmission control strategy about each electric device; the power cut and transmission control strategy is sent to a power grid management and control center in an instruction mode, and corresponding power cut and transmission control instruction operation is carried out. The intelligent degree of power grid management is improved, the reliability and efficiency of the system are enhanced, and the requirements of a modern power grid for flexibility and high efficiency are met.

Description

Technical Field

[0001] The present invention relates to the technical field of smart grid management and control, and in particular to an intelligent remote power outage control method based on the Internet of Things. Background Art

[0002] At present, traditional power grids are evolving towards smart grids, in which real-time communication and control are crucial to optimizing energy distribution and enhancing system resilience. With the development of big data technology, machine learning and artificial intelligence algorithms are widely used to process large amounts of data generated by IoT (Internet of Things) devices, realizing analytical processing and automatic decision-making in power system operations.

[0003] However, existing technologies still have many limitations in effectively managing the dynamic complexity of modern power grids. Current load forecasting methods usually rely on linear models or traditional statistical techniques, which cannot fully capture the complex nonlinear relationship between load changes and electricity demand, which is affected by multiple factors such as weather conditions, user behavior and economic activities. This leads to inaccurate predictions, which in turn triggers suboptimal control strategies, resulting in overloads, increased operating costs and even power outages.

[0004] In addition, traditional control models usually adopt a single-input single-output approach, which cannot fully handle the multi-variable and multi-objective characteristics of power grid operations, resulting in inefficient power distribution and challenges in implementing effective demand response plans. There is a lack of integration between predictive analysis and control systems. Even if accurate prediction results are obtained, it is difficult to use them in a timely manner to guide real-time adjustments to control strategies. This limitation of being unable to respond quickly to rapid changes in load demand has hindered the improvement of efficiency and reliability of power grid operations. Summary of the invention

[0005] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract of the specification and the title of the invention of this application to avoid blurring the purpose of this section, the abstract of the specification and the title of the invention, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0006] In view of the above existing problems, the present invention is proposed.

[0007] Therefore, the technical problem solved by the present invention is that the existing power grid control technology has the problems of insufficient real-time load prediction accuracy, difficulty in handling the nonlinear relationship of load demand, and inefficient power outage control strategy, which leads to reduced power supply reliability, increased operating costs and decreased customer satisfaction.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions: using the Internet of Things technology to obtain the load changes and power demand of each power-consuming device in the power grid in real time;

[0009] Using a weighted optimized extreme learning machine, a nonlinear mapping relationship between the load change and the power demand is established, and a prediction result of future load demand is output;

[0010] The prediction results are input into a control model constructed based on a multi-input multi-output fuzzy control algorithm to obtain a power outage control strategy for each power-consuming device;

[0011] The power outage control strategy is sent to the power grid control center in the form of instructions, and the corresponding power outage control instruction operation is performed.

[0012] As a preferred solution of the intelligent remote power outage control method based on the Internet of Things described in the present invention, the load change data, power demand data, environment and time parameters of each power-consuming device are collected in real time according to the set sampling frequency through the Internet of Things sensor, and the sensor sends the collected data to the Internet of Things gateway through the communication module. The Internet of Things gateway performs preliminary processing on the collected data and forwards it to the cloud server and the data center of the power grid control center, which receives and stores the data from each power-consuming device;

[0013] The set sampling frequency is once per second.

[0014] As a preferred solution of the intelligent remote power outage control method based on the Internet of Things described in the present invention, the load change data at least includes active power, reactive power, apparent power and load change rate;

[0015] The electricity demand data is historical electricity consumption;

[0016] The environmental and time parameters include temperature, humidity and time characteristics, and the time characteristics include hours, weeks and months.

[0017] As a preferred solution of the intelligent remote power outage control method based on the Internet of Things described in the present invention, the collected load change data, the power demand data, the environment and time parameters are subjected to data cleaning and time alignment processing, and then feature extraction is performed to construct a feature vector, including:

[0018] X(t)=[P(t),ΔP(t),E hist (t),D(t),T(t),H(t),t hour ,t day ,t month ]

[0019] Among them, P(t) is the current active power, ΔP(t) is the change in active power, E hist(t) is the historical electricity consumption, D(t) is the predicted electricity demand, T(t) and H(t) are the ambient temperature and humidity respectively, t hour ,t day ,t month The columns are hour, day of the week, and month.

[0020] As a preferred solution of the intelligent remote power outage control method based on the Internet of Things described in the present invention, the output of the prediction result of future load demand includes:

[0021]

[0022] in, To predict future load demand, is the hidden layer output vector h new is the transpose of , and β is the output weight.

[0023] As a preferred solution of the intelligent remote power outage control method based on the Internet of Things described in the present invention, the prediction result is input into a control model constructed based on a multi-input multi-output fuzzy control algorithm to obtain a power outage control strategy for each power-consuming device, including:

[0024] For each input variable, calculate its membership under each fuzzy set according to its actual value and membership function;

[0025] Then calculate the activation of each rule;

[0026] According to the activation degree, weighted synthesis is performed on the fuzzy set of the output variable;

[0027] The value of the output variable is calculated using the centroid method, and its mathematical expression formula is:

[0028]

[0029] Where F is the fuzzy control function, which is determined by the fuzzy rule base and defuzzification method. To predict the load demand, P(t) is the current load level, U i is the priority of the ith electrical equipment. The higher the value, the more important the equipment is. The value range is [1,5]. T(t) is the ambient temperature at the current moment. C i (t) is the power outage control strategy value.

[0030] As a preferred solution of the intelligent remote power outage control method based on the Internet of Things described in the present invention, the power outage control strategy value C i The value range of (t) is [0,1], which is the control strategy value of the i-th electrical equipment, where:

[0031] When Ci When (t) = 0, there is a power outage;

[0032] When 0 <C i When (t)<1, partial power supply or load regulation occurs;

[0033] When C i (t) = 1, then the power supply is normal.

[0034] As a preferred solution of the intelligent remote power outage control method based on the Internet of Things described in the present invention, the power outage control strategy specifically includes:

[0035] When both the predicted load demand and the current load level are high and the equipment priority is low, a power outage is taken;

[0036] When the predicted load demand and current load level are moderate, and the equipment priority is medium, partial power supply or load adjustment measures are taken;

[0037] Maintain normal power supply when predicted load demand and current load levels are low, or when equipment priority is high.

[0038] As a preferred solution of the intelligent remote power outage control method based on the Internet of Things described in the present invention, it also includes:

[0039] Set the continuous power outage control strategy value C i (t) Discretization to generate clear control actions;

[0040] Calculate control parameters, including power limits;

[0041] Encapsulate control instructions in a unified data format;

[0042] Sending the control command to the power grid control center through the power dedicated network;

[0043] The control center decrypts and verifies the control instruction, parses the instruction content, and prepares to execute:

[0044] A control command is issued to a specific device, and after the device executes the corresponding power-off operation, the execution result is fed back, and the control center updates the status.

[0045] Beneficial effects of the present invention:

[0046] 1. Provide high-quality data foundation for subsequent load forecasting and control strategies, and improve data accuracy and real-time performance;

[0047] 2. By using the rapid learning ability and weighted optimization method of the extreme learning machine, the model's ability to capture nonlinear relationships and prediction accuracy are improved, overcoming the limitation of traditional prediction methods in insufficient processing of nonlinear features, providing more accurate future load demand predictions, improving prediction accuracy and reducing prediction errors;

[0048] 3. Effectively handle uncertainty and ambiguity, provide flexible and robust control solutions, optimize power distribution, improve grid operation efficiency, and meet the power needs of different devices and users;

[0049] 4. Ensure the accurate transmission and timely execution of control instructions, ensure the safe and stable operation of the power grid, improve the control response speed, reduce operational risks, and enhance the intelligent management level of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0051] Figure 1 It is a flow chart of the intelligent remote power outage and power transmission control method based on the Internet of Things shown in the present invention;

[0052] Figure 2 A comparison diagram of load prediction results shown in the present invention;

[0053] Figure 3 A comparison diagram of power grid losses shown in the present invention;

[0054] Figure 4 This is a power supply reliability comparison diagram shown in the present invention. DETAILED DESCRIPTION

[0055] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0056] Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without making any creative work should fall within the scope of protection of the present invention.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0058] According to an embodiment of the present invention, Figure 1 The flowchart shown is an intelligent remote power outage control method based on the Internet of Things, which specifically includes the following steps:

[0059] S1. Use the Internet of Things technology to obtain the load changes and power demand of each power-consuming device in the power grid in real time. Among them, what needs to be explained in this step is:

[0060] The IoT sensor collects the load change data, power demand data, environment and time parameters of each power-consuming device in real time according to the set sampling frequency (such as once per second). The sensor sends the collected data to the IoT gateway through the communication module. The IoT gateway performs preliminary processing on the data and forwards it to the cloud server or the data center of the power grid control center, which receives and stores the data from each power-consuming device and uses a database (such as MySQL, MongoDB, HBase) for secure data storage; wherein:

[0061] (1) Load change data

[0062] Active power:

[0063] P(t)=V(t)×I(t)×cosφ(t)

[0064] Where P(t) is active power, V(t) is voltage, I(t) is current, and cosφ(t) is power factor, which has no unit.

[0065] Reactive power:

[0066] Q(t)=V(t)×I(t)×sinφ(t)

[0067] Among them, Q(t) is the reactive power, sinφ(t) is the sine value of the power factor angle;

[0068] Apparent power:

[0069] S(t)=V(t)×I(t)

[0070] Where, S(t) is the apparent power;

[0071] Load change rate ΔP(t):

[0072] ΔP(t)=P(t)-P(t-Δt)

[0073] Among them, ΔP(t) is the change in active power, Δt is the time interval;

[0074] (2) Electricity demand data

[0075] Historical electricity consumption E hist (t):

[0076]

[0077] Among them, E hist (t) is the cumulative power consumption in the time period [tT, t], T is the historical time window, and τ is the integral variable, indicating the time interval;

[0078] (3) Environmental and time parameters

[0079] Temperature T(t): unit is Celsius (℃);

[0080] Humidity H(t): unit is percentage (%);

[0081] Time feature t: includes hours, weeks, and months, reflecting the periodicity of time;

[0082] Furthermore, the collected data is cleaned and time-aligned, and then feature extraction is performed to construct the feature vector X(t):

[0083] X(t)=[P(t),ΔP(t),E hist (t),D(t),T(t),H(t),t hour ,t day ,t month ]

[0084] Among them, P(t) is the current active power, ΔP(t) is the change in active power, E hist (t) is the historical electricity consumption, D(t) is the predicted electricity demand, T(t) and H(t) are the ambient temperature and humidity respectively, t hour ,t day ,t month for time features, such as hour, day of the week, month;

[0085] The feature vector X(t) and the corresponding target variable (such as the load demand Y(t+τ) at the future moment) form a data set D:

[0086] D={(X(t),Y(t+τ))}

[0087] Among them, Y(t+τ) is the load demand at the future time t+τ, that is, the target variable, and τ is the predicted time interval;

[0088] As an example, use CSV files, Excel tables, or database tables for storage to facilitate subsequent model calls.

[0089] S2. Use the weighted optimized extreme learning machine to establish a nonlinear mapping relationship between load changes and power demand, and output the prediction results of future load demand. Among them, what needs to be explained in this step is:

[0090] The feature vector X(t) obtained in step S1 is used as the input for prediction training in step S2; its mathematical expression formula is:

[0091] X(t)=[P(t),ΔP(t),E hist (t),D(t),T(t),H(t),t hour ,t day ,t month ]

[0092] Among them, P(t) is the active power at the current moment, ΔP(t) is the change in active power, and E hist (t) is the historical electricity consumption, D(t) is the planned electricity demand value, T(t) is the ambient temperature, H(t) is the ambient humidity, t hour is the current hour, ranging from 0 to 23, t day The current day of the week, ranging from 1 to 7, t month The current month, ranging from 1 to 12;

[0093] Define the target variable Y(t) as the load demand at the future moment output by the model, i.e. the value to be predicted;

[0094] In an optional embodiment, the extreme learning machine includes the following structure:

[0095] Input layer: The input feature vector dimension is n, that is

[0096] Hidden layer: The number of hidden nodes is L;

[0097] Input weight matrix

[0098]

[0099] Among them, w ij is the weight between the i-th hidden node and the j-th input node, which is randomly generated;

[0100] Bias Vector

[0101] b=[b1,b2,…,b L ] T

[0102] Among them, b L is the bias of the Lth hidden node, randomly generated;

[0103] Activation function g(·): such as Sigmoid function and ReLU function;

[0104] Output layer: output weight vector

[0105] β=[β1,β2,…,β L ] T

[0106] Train it to further improve the prediction accuracy, specifically:

[0107] Initialize the parameters and randomly generate the input weight matrix W and bias vector b;

[0108] Calculate the hidden layer output matrix H;

[0109] For each sample X k (t), calculate the hidden layer output:

[0110]

[0111] Among them, h i (X k ) is the output of the kth sample at the i-th hidden node, x kj is the j-th eigenvalue of the k-th sample;

[0112] Construct the hidden layer output matrix

[0113]

[0114] Where N is the number of samples;

[0115] Assign different weights to different samples to form a diagonal weight matrix

[0116] W d =diag(w1,w2,…,w N )

[0117] Among them, w k is the weight of the kth sample, which is set according to the importance of the sample or the size of the error;

[0118] The weighted least squares method is used to solve the output weight β, and the objective function is:

[0119]

[0120] is the target variable vector, Y=[Y1,Y2,…,Y N ] T ;

[0121] Solution formula:

[0122] β=(H T W d H) -1 H T W d Y

[0123] Among them, H T is the transpose of the hidden layer output matrix, (H T W d H) -1 is the inverse of the matrix. If it is not invertible, a regularization method can be used, such as adding a regularization term λ:

[0124] β=(H T W d H+λI) -1 H T W d Y

[0125] Where λ is the regularization coefficient, a non-negative real number, and I is the L×L unit matrix;

[0126] Going further, predictions are made for new samples, including:

[0127] For the new input feature vector X obtained in real time new , calculate the hidden layer output:

[0128]

[0129] Among them, x new,j is the jth eigenvalue of the new sample;

[0130] Construct hidden layer output vector

[0131] h new =[h1(X new ),h2(X new ),…,h L (X new )]

[0132] Predict future load demand:

[0133]

[0134] in, For predicted future load demand;

[0135] Will It is input as input variables into the fuzzy control model of step S3;

[0136] In an optional implementation, corresponding to the future time t+τ (τ is the prediction time interval), the time point is predicted to obtain a series of predicted values, which are displayed in a list or array form, such as

[0137] As an example, define the current time t = 10:00, characteristic data P(t) = 1000W, ΔP(t) = 50W, E hist (t)=20kWh, D(t)=22kWh, T(t)=28℃, H(t)=70%, t hour =10, t day =3(Wednesday), t month =6(June);

[0138] Feature vector X(t) = [1000, 50, 20, 22, 28, 70, 10, 3, 6];

[0139] The number of hidden layer nodes L = 10;

[0140] Randomly initialize parameters W and b;

[0141] The activation function uses the Sigmoid function;

[0142] Calculate the hidden layer output vector:

[0143] h new =[h1(X(t)),h2(X(t)),…,h 10 (X(t))]

[0144] Using the trained output weight β, predict the load demand:

[0145]

[0146] It is predicted that the load demand at a future time (such as t+1 hour, i.e. 11:00) will be 1050W.

[0147] S3, input the prediction results into the control model constructed based on the multi-input multi-output fuzzy control algorithm to obtain the power outage control strategy for each power-consuming device. Among them, what needs to be explained in this step is:

[0148] Defining input variables includes:

[0149] Forecasting load demand The load demand forecast value at the future time t obtained from step S2;

[0150] The current load level P(t), the active power at the current moment obtained from step S1;

[0151] Device Priority U i , the value range is [1,5], the priority of the i-th electrical equipment, the higher the value, the more important the equipment;

[0152] Ambient temperature T(t), the ambient temperature at the current moment;

[0153] Defining output variables includes:

[0154] Power outage control strategy C i (t), the value range is [0,1], the control strategy value for the i-th electrical equipment, the specific meaning is as follows:

[0155] C i (t) = 0: power failure;

[0156] 0 <C i (t)<1: partial power supply or load regulation;

[0157] C i (t) = 1: normal power supply;

[0158] Specifically, the implementation process of this step includes the following steps:

[0159] Fuzzify the input variables and output variables and define the corresponding fuzzy sets. For example, the fuzzy sets of input variables include:

[0160] Forecasting load demand Fuzzy set of: Low, Medium, High;

[0161] The fuzzy set of the current load level P(t): Low, Medium, High;

[0162] Device Priority U i Fuzzy set: Low (Low Priority), Medium (Medium Priority), High (HighPriority);

[0163] Fuzzy set of ambient temperature T(t): Low Temp, Medium Temp, High Temp;

[0164] For example, the fuzzy set of output variables includes:

[0165] Control Strategy C i Fuzzy set of (t): power outage (Off), partial power supply (Partial), normal power supply (On);

[0166] The membership function adopts a triangular or trapezoidal membership function, and a suitable membership function can be adaptively adopted according to the actual application situation, and is not limited to it in this embodiment;

[0167] In this implementation, we take load demand prediction as an example, and the corresponding fuzzy sets are:

[0168] Low:

[0169]

[0170] Medium:

[0171]

[0172] High:

[0173]

[0174] Among them, Y low1 , Y low2 , Y med1 , Y med2 , Y med3 , Y high1 , Y high2 It is the cutoff point of the membership function and is set according to the actual load demand range;

[0175] Furthermore, a fuzzy rule base is established based on the fuzzy set of input variables. The form of the rules is:

[0176] Rule 1: If the predicted load demand is "high" and the current load level is "high" and the equipment priority is "low", then the control strategy is "power outage";

[0177] Rule 2: If the predicted load demand is "low" and the current load level is "low" and the equipment priority is "high", then the control strategy is "normal power supply";

[0178] Rule 3: If the predicted load demand is "medium" and the current load level is "medium" and the equipment priority is "medium", then the control strategy is "partial power supply";

[0179] Using fuzzy logic reasoning methods (such as Mamdani reasoning), fuzzy operations are performed on input variables to obtain the fuzzy set of output variables, where:

[0180] For each rule, calculate its activation:

[0181]

[0182] Among them, α k is the activation degree of the kth rule, To predict the membership of the fuzzy set corresponding to the load demand in rule k, is the membership degree of the current load level corresponding to the fuzzy set in rule k, is the membership degree of the fuzzy set corresponding to the device priority in rule k, is the membership degree of the fuzzy set corresponding to the ambient temperature in rule k;

[0183] According to the activation degree of all rules, the fuzzy set of the synthesized output is obtained;

[0184] The output fuzzy set obtained by fuzzy reasoning is converted into a specific value using the centroid method, that is, the control strategy value C i (t):

[0185]

[0186] Among them, C i (t) is the control strategy value of the ith device, ranging from [0,1], is the membership function of the output variable, c is the variable of the control strategy value, and Ω is the domain of the control strategy value, i.e. [0,1];

[0187] It is also necessary to explain in detail that the power outage control strategy includes:

[0188] C i (t) = 0: indicates that the power supply to the i-th electrical equipment is immediately cut off;

[0189] 0 <C i (t)<1: indicates that partial power supply or load adjustment is performed on the i-th electrical equipment, such as reducing the power supply voltage or limiting the maximum power;

[0190] C i (t) = 1: indicates that the power supply to the i-th electrical equipment is normal and no adjustment is required;

[0191] For each input variable, calculate its membership under each fuzzy set according to its actual value and membership function;

[0192] Using the above formula α k , calculate the activation of each rule;

[0193] According to the activation degree, the fuzzy set of the output variable is weighted and synthesized;

[0194] The centroid method is used to calculate the specific value of the output variable. The mathematical expression formula is:

[0195]

[0196] Among them, F is the fuzzy control function, which is determined by the fuzzy rule base and defuzzification method;

[0197] Specifically, the power outage control strategy is as follows:

[0198] When both the predicted load demand and the current load level are high and the equipment priority is low, a power outage is taken;

[0199] When the predicted load demand and current load level are moderate, and the equipment priority is medium, partial power supply or load adjustment measures are taken;

[0200] Maintain normal power supply when predicted load demand and current load levels are low, or when equipment priority is high.

[0201] S4. The power outage control strategy is sent to the power grid control center in the form of instructions, and the corresponding power outage control instruction operation is performed. Among them, it is also necessary to explain that:

[0202] The continuous control strategy value C i (t) Discretization to generate clear control actions;

[0203] Calculate control parameters (such as power limits);

[0204] Encapsulate control instructions in a unified data format;

[0205] Use encryption algorithms and digital signatures to ensure the security of instructions;

[0206] Send the command to the power grid control center through the power dedicated network;

[0207] The control center decrypts and verifies the command, parses the command content, and prepares to execute it:

[0208] Send control commands to specific devices. After the devices execute the corresponding power-off operation, they will feedback the execution results and the control center will update the status.

[0209] It should be noted that due to the control strategy C i The value range of (t) is continuous [0,1] and needs to be discretized in order to generate clear control instructions.

[0210] Discretization rules:

[0211]

[0212] Exemplarily, the control parameters (for partial power supply) are calculated, such as power limit:

[0213] P limit,i =C i (t)×P max,i

[0214] Among them, P limit,iis the power limit of the ith device, P max,i is the maximum rated power of the ith device, C i (t) is the control strategy value, ranging from [0,1];

[0215] Furthermore, the control strategy is converted into control instructions in a unified format, including:

[0216] Device ID: uniquely identifies each power-consuming device, such as device number or MAC address;

[0217] Control action (Action): indicates the operation to be performed, such as "power outage", "partial power supply", and "normal power supply";

[0218] Control Parameters: For partial power supply, including power limit parameters;

[0219] Timestamp: the time when the instruction was generated, in the format of YYYY-MM-DD HH:MM:SS;

[0220] Execution Time: The time at which the instruction is required to be executed, such as immediate execution or a specified future time;

[0221] Digital Signature: An encrypted signature used to verify the integrity and source of an instruction;

[0222] Instruction example:

[0223] {

[0224] "DeviceID":"Device_001",

[0225] "Action":"Partial_Power",

[0226] "Parameters":{

[0227] "PowerLimit":5000

[0228] },

[0229] "Timestamp":"2023-10-01 10:00:00",

[0230] "ExecutionTime":"2023-10-01 10:05:00",

[0231] "DigitalSignature":"abcdef1234567890"

[0232] }

[0233] Furthermore, the instruction sending includes the following process:

[0234] Generate instructions according to the control strategy and encapsulate them into the specified data format;

[0235] Use encryption algorithm to encrypt instruction content and add digital signature;

[0236] Sending instructions to the power grid control center via the communication network;

[0237] The power grid control center is equipped with a dedicated command receiving module to monitor the arrival of commands in real time;

[0238] Decrypt the command using a symmetric key or the recipient's private key;

[0239] Use the sender's public key to verify the digital signature of the command to ensure that the command has not been tampered with;

[0240] Check whether the format of the instruction complies with the predefined protocol and data format;

[0241] Extract device identifier, control action, control parameter, and execution time information from the instruction;

[0242] Formulate a specific execution plan based on the execution time of the instruction and the operation status of the power grid;

[0243] Send the parsed control commands to the corresponding execution devices, such as circuit breakers and intelligent switches;

[0244] After completing the operation, the execution device will feedback the execution result and device status to the control center.

[0245] As an example, define the DeviceID of device A: "Device 001 ", control strategy value C A (t) = 0: power outage is required, maximum rated power P max,A =10000W, Device ID of device B: "Device 002 ", control strategy value C B (t) = 0.5: Partial power supply, maximum rated power P max,B =8000W,

[0246] The instruction content of device A is:

[0247] {

[0248] "DeviceID":"Device_001",

[0249] "Action":"Power_Off",

[0250] "Parameters":{},

[0251] "Timestamp":"2023-10-01 10:00:00",

[0252] "ExecutionTime":"2023-10-01 10:05:00",

[0253] "DigitalSignature":"signature_device_A"

[0254] }

[0255] Device B's instructions:

[0256] Calculate the power limit:

[0257] P limit,B =C B (t)×P max,B =0.5×8000=4000W

[0258] The instruction content is:

[0259] {

[0260] "DeviceID":"Device_002",

[0261] "Action":"Partial_Power",

[0262] "Parameters":{

[0263] "PowerLimit":4000

[0264] },

[0265] "Timestamp":"2023-10-01 10:00:00",

[0266] "ExecutionTime":"2023-10-01 10:05:00",

[0267] "DigitalSignature":"signature_device_B"

[0268] }

[0269] Use AES to encrypt instruction content and use RSA to digitally sign instructions;

[0270] The encrypted instructions are sent to the receiving module of the power grid control center through the power network;

[0271] The control center uses the AES key to decrypt the command;

[0272] Verify the digital signature using the sender's public key;

[0273] Parsing instruction content includes extracting device ID, control action, and parameters;

[0274] Device A performs the following operations:

[0275] Control action: power failure;

[0276] Execution process: Send a power-off command to device A, device A receives the command, performs the power-off operation, and device A feedbacks the execution result;

[0277] Device B performs the following operations:

[0278] Control action: partial power supply;

[0279] Control parameters: Power limit is 4000W;

[0280] Execution process: A power limit command is sent to device B, which adjusts the power supply to 4000 W. Device B then feeds back the execution result.

[0281] In order to verify the effectiveness of an intelligent remote power outage control method based on the Internet of Things proposed in an embodiment of the present invention, a comparative experiment is conducted with a traditional single-input single-output (SISO) control method. The traditional method cannot fully handle the multi-variable and multi-objective characteristics in power grid operation, resulting in poor control effect. The following comparative experiments are designed to verify the advantages of the method of the present invention in terms of load prediction accuracy and power outage control efficiency, wherein:

[0282]

Experimental subjects

[0283] This experiment selected a typical distribution network area of ​​a city, covering residential, commercial and industrial power loads, and selected five major power-consuming devices, which were numbered as follows:

[0284] Equipment 1 (residential area): The main power supply target is residential users. The power load is relatively stable, but it fluctuates greatly during peak hours in the morning and evening.

[0285] Equipment 2 (commercial center): The power supply targets commercial users in shopping malls and office buildings, and the power load is high during working hours;

[0286] Equipment 3 (industrial park): The power supply target is manufacturing enterprises, with large power load and long duration;

[0287] Equipment 4 (hospital): a key power user with extremely high requirements for power supply reliability;

[0288] Equipment 5 (school): The power load is higher during the day and lower at night;

[0289] These devices have different power consumption characteristics and load change patterns, which can fully reflect the diversity and complexity of the power grid;

[0290]

Experimental environment conditions

[0291] Sensor equipment: Install voltage sensors, current sensors and power metering devices at key nodes of each electrical equipment with an accuracy of 0.1%;

[0292] IoT communication module: Uses communication modules that support LoRa and 4G networks to achieve real-time data transmission;

[0293] Server equipment: Configure high-performance servers for data storage and processing, with a CPU frequency of 3.0 GHz and a memory of 32 GB;

[0294] Operating system: The server runs the Linux operating system;

[0295] Data processing platform: Use Hadoop big data platform and combine it with MySQL database for data management;

[0296] Algorithm implementation tools: Use MATLAB R2021a to develop and implement the algorithm, and use Python to perform some data preprocessing;

[0297]

Experiment time

[0298] Data collection period: one week, sampling interval is 5 minutes;

[0299] Test time period: Select typical weekdays and weekends for testing, covering peak and valley load periods;

[0300]

Experimental process

[0301] Traditional method:

[0302] Install basic voltage and current sensors to collect historical load data of each device;

[0303] Use a simple linear regression model to make forecasts based only on historical load data;

[0304] A single-input single-output control model is used to formulate a power outage strategy (multiple influencing factors cannot be considered);

[0305] Send the control strategy to the power grid control center to execute the corresponding power outage operation;

[0306] The method of the present invention:

[0307] Using IoT technology, a variety of data can be collected in real time, including load changes, ambient temperature and humidity, and time parameters;

[0308] Use weighted optimized extreme learning machine to establish nonlinear mapping relationship between load change and power demand;

[0309] The prediction results are input into the multi-input multi-output fuzzy control model, and various factors are comprehensively considered to formulate an optimized power outage control strategy;

[0310] Send the control strategy in the form of instructions to the power grid control center to perform corresponding power outage control operations;

[0311] Reference Figure 2 ,The load prediction curve (black dotted line) of the method of the present invention is almost completely coincident with the actual load curve (black solid line), and the prediction error is extremely small, while the prediction curve (black dashed line) of the traditional method has a significant deviation from the actual load, especially in the period of drastic load changes (such as morning peak and evening peak);

[0312] Reference Figure 3 , on all devices, the grid loss of the method of the present invention is lower than that of the traditional method. Since the method of the present invention optimizes power dispatching and reduces unnecessary power loss through accurate load prediction and intelligent power outage control strategy, the average grid loss of the method of the present invention is reduced by 34.5% compared with the traditional method;

[0313] Reference Figure 4 The method of the present invention is significantly superior to the traditional method in terms of power supply reliability. The traditional method cannot respond quickly to load changes, which easily leads to equipment overload or insufficient power supply, reducing power supply reliability. The method of the present invention can adjust the power supply strategy in time, avoid overload and power supply interruption, and improve the continuity and stability of power supply. The average power supply reliability of the method of the present invention is improved by 3.2%;

[0314] Preferably, the method of the present invention is superior to the traditional method in all indicators, reflecting its advantages in dealing with multi-variable and multi-objective power grid control problems. Through the application of Internet of Things technology, real-time data collection and transmission are realized, providing data guarantee for the accuracy of the model. At the same time, a weighted optimized extreme learning machine and a multi-input and multi-output fuzzy control algorithm are used to construct an efficient prediction and control model, thereby improving the intelligence level of the power grid.

[0315] Preferably, the experimental results fully verify the effectiveness and superiority of the method of the present invention. Compared with the traditional single-input single-output control method, the method of the present invention can more accurately predict load demand, formulate a more optimized power outage control strategy, significantly reduce grid losses, and improve power supply reliability.

[0316] The aforementioned method for preprocessing and feature extraction of collected data can be performed using methods and means in the prior art, and will not be described in detail in this example.

[0317] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent remote power outage control method based on the Internet of Things, characterized in that: include: Use IoT technology to obtain real-time load changes and power demand of each power-consuming device in the power grid; Using a weighted optimized extreme learning machine, a nonlinear mapping relationship between the load change and the power demand is established, and a prediction result of future load demand is output; The prediction results are input into a control model constructed based on a multi-input multi-output fuzzy control algorithm to obtain a power outage control strategy for each power-consuming device; The power outage control strategy is sent to the power grid control center in the form of instructions, and the corresponding power outage control instruction operation is performed.

2. The intelligent remote power outage control method based on the Internet of Things according to claim 1 is characterized in that: The IoT sensor collects load change data, power demand data, environment and time parameters of each power-consuming device in real time according to the set sampling frequency. The sensor sends the collected data to the IoT gateway through the communication module. The IoT gateway performs preliminary processing on the collected data and forwards it to the cloud server and the data center of the power grid control center, which receives and stores data from each power-consuming device; The set sampling frequency is once per second.

3. The intelligent remote power outage control method based on the Internet of Things according to claim 2 is characterized in that: The load change data at least includes active power, reactive power, apparent power and load change rate; The electricity demand data is historical electricity consumption; The environmental and time parameters include temperature, humidity and time characteristics, and the time characteristics include hours, weeks and months.

4. The intelligent remote power outage control method based on the Internet of Things according to claim 3 is characterized in that: The collected load change data, the power demand data, the environment and time parameters are cleaned and time-aligned, and then feature extracted to construct a feature vector, including: X(t)=[P(t),ΔP(t),E hist (t),D(t),T(t),H(t),t hour ,t day ,t month ] Among them, P(t) is the current active power, ΔP(t) is the change in active power, E hist (t) is the historical electricity consumption, D(t) is the predicted electricity demand, T(t) and H(t) are the ambient temperature and humidity respectively, t hour ,t day ,t month The hours, days of the week, and months respectively.

5. The intelligent remote power outage control method based on the Internet of Things according to claim 1 is characterized in that: The output is a prediction result of future load demand, including: in, To predict future load demand, is the hidden layer output vector h new is the transpose of , and β is the output weight.

6. The intelligent remote power outage control method based on the Internet of Things according to claim 5 is characterized in that: The prediction results are input into a control model constructed based on a multi-input multi-output fuzzy control algorithm to obtain a power outage control strategy for each power-consuming device, including: For each input variable, calculate its membership under each fuzzy set according to its actual value and membership function; Then calculate the activation of each rule; According to the activation degree, weighted synthesis is performed on the fuzzy set of the output variable; The value of the output variable is calculated using the centroid method, and its mathematical expression formula is: Where F is the fuzzy control function, which is determined by the fuzzy rule base and defuzzification method. To predict the load demand, P(t) is the current load level, U i is the priority of the ith electrical equipment. The higher the value, the more important the equipment is. The value range is [1,5]. T(t) is the ambient temperature at the current moment. C i (t) is the power outage control strategy value.

7. The intelligent remote power outage control method based on the Internet of Things according to claim 6 is characterized in that: The power outage control strategy value C i The value range of (t) is [0,1], which is the control strategy value of the i-th electrical equipment, where: When C i When (t) = 0, there is a power outage; When 0 <C i When (t)<1, partial power supply or load regulation occurs; When C i (t) = 1, then the power supply is normal.

8. The intelligent remote power outage control method based on the Internet of Things according to claim 6 is characterized in that: The power outage control strategy specifically includes: When both the predicted load demand and the current load level are high and the equipment priority is low, a power outage is taken; When the predicted load demand and current load level are moderate, and the equipment priority is medium, partial power supply or load adjustment measures are taken; Maintain normal power supply when predicted load demand and current load levels are low, or when equipment priority is high.

9. The intelligent remote power outage control method based on the Internet of Things according to claim 6 is characterized in that: Also includes: Set the continuous power outage control strategy value C i (t) Discretization to generate clear control actions; Calculate control parameters, including power limits; Encapsulate control instructions in a unified data format; Sending the control command to the power grid control center through the power dedicated network; The control center decrypts and verifies the control instruction, parses the instruction content, and prepares to execute: A control command is issued to a specific device, and after the device executes the corresponding power-off operation, the execution result is fed back, and the control center updates the status.

Images