Intelligent power distribution network energy dynamic scheduling management system and method

Abstract

The invention discloses an intelligent power distribution network energy dynamic scheduling management system and method, and relates to the technical field of intelligent power distribution networks, and the system comprises the following components: a data storage module, a data analysis module, a scheduling decision module, a communication interface module, a plurality of regional power distribution subsystems, and a data collection and transmission module. According to the invention, a data storage module, a data analysis module, a scheduling decision module, a communication interface module, a regional power distribution subsystem module and a data acquisition and transmission module are integrated, so that the dynamic scheduling management of the energy of the intelligent power distribution network is realized; the integrated management system can collect and analyze voltage, current, power and load data from each regional power distribution network and power generation data of distributed energy in real time, comprehensive and accurate information support is provided for scheduling decisions, and meanwhile, a multi-target optimization scheduling algorithm adopted by the system can improve the scheduling efficiency. Multiple targets of resource optimization configuration, operation efficiency improvement and reliability enhancement can be comprehensively considered, and an optimal scheduling scheme is generated.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent distribution networks, and particularly to an intelligent distribution network energy dynamic scheduling management system and method. Background Art

[0002] With the rapid development of smart grid technology, as an important part of the power system, the energy scheduling management of intelligent distribution networks faces more and more challenges. The intelligent distribution network needs to collect and process a large amount of data from regional distribution networks in real time, including voltage, current, power, load data, and power generation data of distributed energy sources, to ensure the safe, stable, and efficient operation of the power grid.

[0003] Traditional technologies have deficiencies. Traditional distribution network scheduling methods are often limited to a single region and lack information sharing and collaborative operation mechanisms between different regions, which leads to many problems in energy allocation and scheduling, such as the inability to optimize the configuration of resources, the difficulty in realizing energy complementarity between regions, and the limitation of the operation efficiency and reliability of the entire regional power grid.

[0004] In the face of complex and changing energy demands and the large-scale access of distributed energy sources, traditional scheduling methods cannot process a large amount of data in real time and accurately, are difficult to rationally allocate energy through intelligent algorithms, and cannot meet the requirements of the cross-regional collaborative development of modern intelligent distribution networks.

[0005] In summary, there are obvious deficiencies in traditional scheduling methods and they cannot meet the requirements of modern intelligent distribution networks for energy scheduling management. Therefore, it is particularly important to develop an intelligent distribution network energy dynamic scheduling management system and method. Summary of the Invention

[0006] The purpose of the present invention is to make up for the deficiencies of the existing technology and provide an intelligent distribution network energy dynamic scheduling management system and method. It can realize the real-time, efficient, and dynamic scheduling management of the energy of the intelligent distribution network by integrating multiple modules such as data storage, data analysis, scheduling decision-making, communication interfaces, regional distribution sub-systems, and data collection and transmission. This integrated management system can not only improve the accuracy and reliability of data, but also generate the optimal scheduling plan through multi-objective optimization scheduling algorithms, thereby comprehensively improving the energy utilization efficiency and overall operation performance of the power grid.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: an intelligent distribution network energy dynamic scheduling management system and method, and the system includes the following components: a data storage module, a data analysis module, a scheduling decision-making module, a communication interface module, multiple regional distribution sub-systems, and a data collection and transmission module;

[0008] The data storage module: stores real-time data from each regional power distribution subsystem and historical dispatching data. The real-time data includes voltage V, current I, power P, and load L data of each regional power grid, power generation data E of distributed energy sources, and power consumption demand data D of users. The historical dispatching data includes energy distribution plans and dispatching execution result information in different past time periods;

[0009] The data analysis module: uses big data analysis algorithms to deeply mine and analyze the collected data, and uses the Kalman filtering algorithm to denoise the real-time data. The formula is:

[0010]

[0011] where, is the state prediction value at time k based on the information at time k - 1, A is the state transition matrix, is the state estimate value at time k - 1, B is the control input matrix, u k-1 is the control input at time k - 1, P k|k-1 is the prediction error covariance at time k based on the information at time k - 1, Q is the process noise covariance, K k is the Kalman gain, H is the observation matrix, Z k is the observation value at time k, R is the observation noise covariance, is the state estimate value at time k, P k|k is the estimation error covariance at time k. By analyzing the historical data and the denoised real-time data, valuable information such as load prediction and energy distribution is extracted. For load prediction, an artificial neural network algorithm with adaptive weights is used. The input layer nodes correspond to different influencing factors, time t, weather W, and historical load L h , the hidden layer nodes perform nonlinear transformation through the activation function , and the output layer is the predicted load value The network weights w and biases are adaptively adjusted by minimizing the mean square error between the predicted value and the actual value , where N is the number of samples, L n is the actual load value, is the predicted load value. The initial values of the weights are generated by random numbers and then updated according to the backpropagation algorithm during the training process;

[0012] The dispatching decision module: according to the results of the data analysis module, uses a multi-objective optimal dispatching algorithm to generate an optimal dispatching plan. This algorithm takes resource optimal allocation, operation efficiency improvement, and reliability enhancement as multiple objectives, and constructs an objective function:

[0013] F = ω1×C r +ω2×C e +ω3×Crli

[0014] Among them, C r represents the resource allocation cost, and C e represents the operation efficiency index, and C rli represents the reliability index. ω1, ω2, and ω3 are the weights corresponding to the respective targets. During the solution process, the energy supply and demand balance in each region, the transmission capacity of the power grid, and the access constraint conditions of distributed energy are considered, and the solution is carried out through the particle swarm optimization algorithm. In the particle swarm optimization algorithm, the position of a particle represents a scheduling scheme, the velocity determines the moving direction and distance of the particle, and the position update formula of the particle is:

[0015] x id (t + 1) = x id (t) + v id (t + 1)

[0016] v id (t + 1) = ωv id (t) + c1r1(t)(p id -x id (t)) + c2r2(t)(g d -x id (t))

[0017] Among them, x id (t) is the position of the i-th particle in the d-th dimensional space, v i d(t) is its velocity, ω is the inertia weight, c1 and c2 are learning factors, r1(t) and r2(t) are random numbers between [0, 1], p id is the historical optimal position of the particle itself, and g d is the global optimal position. The inertia weight ω is dynamically adjusted with the number of iterations to balance the global search and local search capabilities. The formula is:

[0018]

[0019] Among them, ω max and ω min are the maximum and minimum values of the inertia weight respectively, T max is the maximum number of iterations, and t is the current number of iterations;

[0020] The communication interface module: is responsible for high-speed and stable data interaction with each regional sub-power system;

[0021] The multiple regional power distribution subsystems: are respectively responsible for the operation and management of the power distribution network within their respective regions, including data acquisition and local dispatching functions. Each regional power distribution subsystem communicates with the cross-regional intelligent power distribution network collaborative dispatching platform through the data acquisition and transmission module, uploads the power grid operation data V, I, P, L, E, D of this region to the platform, and receives the dispatching instructions issued by the platform;

[0022] The data acquisition and transmission module: adopts high-precision sensors and high-speed communication technologies to collect the operation data V, I, P, L of each regional power distribution network in real time, and transmits these data accurately and error-free to the cross-regional intelligent power distribution network collaborative dispatching platform. The data acquisition frequency f can be dynamically adjusted according to actual needs.

[0023] Furthermore, the data storage module adopts a distributed storage architecture. The data storage module consists of multiple storage nodes, which are distributed in different geographical locations and connected through a high-speed network. The distributed hash table technology is used to store and manage the data. Each storage node is responsible for storing a part of the data. When new data needs to be stored, the storage location is calculated through the hash function according to the characteristics of the data, and the data is stored in the corresponding storage node. The design of the hash function is:

[0024]

[0025] where key is the identifier of the data, a i is a coefficient, n is the degree of the polynomial, and m is the number of storage nodes. This distributed storage architecture can not only improve the storage capacity of the data, but also ensure the integrity and availability of the data through the data replicas of other nodes when a single storage node fails, providing reliable data support for the data analysis module and the dispatching decision-making module. During the data storage process, different types of data are stored separately to facilitate subsequent data query and call. Real-time data is stored in chronological order, and a storage record is made at regular intervals. Historical data is classified and archived according to different dispatching cycles and regional dimensions, facilitating the comparison and analysis of historical data by the data analysis module.

[0026] Even further, the data analysis module also has the ability to detect and process abnormal data. During the data acquisition process, due to sensor failures and communication interferences, abnormal data will appear. The data analysis module uses a statistics-based method to detect abnormal data. First, calculate the mean μ and standard deviation σ of the data:

[0027]

[0028] where x iis the data sample, N is the number of samples, a threshold k is set. When the data point x satisfies |x - μ| > kσ, it is determined as abnormal data. For abnormal data, interpolation method is used for processing. According to the values and time intervals of adjacent normal data points, through the linear interpolation formula:

[0029]

[0030] where, (x1, y1) and (x2, y2) are adjacent normal data points, x is the position of the abnormal data point, y is the substituted value obtained by interpolation. When dealing with outliers of different types of data, the above unified detection and processing methods are followed to ensure the consistency and standardization of data processing, improve the accuracy of data analysis, and provide a more reliable basis for subsequent scheduling decisions.

[0031] Furthermore, when generating the scheduling plan, the scheduling decision-making module also considers the safety constraints of the power grid. The safe operation of the power grid is an important prerequisite for scheduling. Therefore, safety constraint conditions are added to the multi-objective optimal scheduling algorithm. Node voltage constraint:

[0032] V min ≤V i ≤V max

[0033] where, V i is the voltage of node i, V min and V max are the lower and upper limits of the node voltage respectively, and these two values are determined according to the design standards and operation requirements of the power grid;

[0034] Line transmission power constraint:

[0035]

[0036] where, P ij is the transmission power of line ij, is the maximum transmission power of line ij, which is determined by the physical parameters and safety standards of the line. The short-circuit current constraint of the system is also considered. Excessive short-circuit current will damage the power grid equipment. Therefore, it is necessary to limit the short-circuit current within a certain range. Let the short-circuit current be I sc , and its constraint condition is Determined according to the tolerance of power grid equipment. When solving the multi-objective optimal scheduling algorithm, these safety constraint conditions are used as limiting conditions to ensure that the generated scheduling plan can ensure the safe and stable operation of the power grid while meeting the energy distribution and scheduling objectives. By continuously optimizing the scheduling plan, on the premise of meeting the safety constraints, the rationality of resource allocation and operation efficiency are improved as much as possible.

[0037] Furthermore, the communication interface module of the tumor communication adopts an encrypted communication technology to ensure the security of data transmission. During the data interaction process, the communication interface module encrypts the transmitted data, using a combination of asymmetric encryption algorithm and symmetric encryption algorithm. Through the RSA algorithm, a public key K pub and a private key K pri are generated. The public key is used to encrypt the data, and the private key is used to decrypt it. Before data transmission, the sender uses the public key K pab of the receiver to encrypt the data, and then sends the encrypted data. After receiving the data, the receiver uses its own private key K pri to decrypt it. To improve the encryption efficiency, for the encryption of a large amount of data, the AES algorithm is adopted, and the symmetric key K sym is used to encrypt and decrypt the data. The symmetric key K sym is encrypted and transmitted through the RSA algorithm to ensure the security of the key. The specific process is as follows: The sender generates the symmetric key K sym , uses the public key K pab of the receiver to encrypt K sym , obtains the encrypted symmetric key K enc,将 K enc , and sends the encrypted K sym and the encrypted data to the receiver together. The receiver uses the private key K pri to decrypt K enc to obtain K sym , and then uses K sym to decrypt the data. During the communication process, digital signature technology is also adopted to ensure the integrity and non-repudiation of the data. The sender performs a hash operation on the data to obtain a hash value H, uses its own private key K pri to sign H to obtain a signature value S, and sends the data, the signature value S, and the public key K pnub to the receiver together. The receiver uses the public key K pub of the sender to verify the signature value S, and performs a hash operation on the received data to compare whether the hash values are the same, so as to ensure the integrity and non-repudiation of the data.

[0038] Furthermore, the regional power distribution subsystem has a local emergency dispatching function. When the communication with the cross-regional intelligent power distribution network collaborative dispatching platform is interrupted or the platform fails, the regional power distribution subsystem can automatically switch to the local emergency dispatching mode. In the local emergency dispatching mode, the regional power distribution subsystem performs energy allocation and dispatching according to the pre-set local dispatching strategy and the historical data stored locally. The local dispatching strategy is formulated based on the analysis of the historical operation data and load characteristics of the local power grid, and gives priority to ensuring the power consumption needs of important users. The determination of important users is comprehensively judged according to the industry attributes of users and the stability factors of power consumption loads. For the load demand L of important usersimp , it is preferentially satisfied during the scheduling process to ensure its power supply reliability. At the same time, the output power of the power generation equipment is adjusted according to the load prediction, and the local historical load data L is used for the load prediction. h , combined with the local weather data W and time factor t, through a simple linear regression model L pred = a0 + a1t + a2W + a3L h for prediction, where a0, a1, a2, and a3 are coefficients obtained by fitting historical data. According to the prediction results, the output power P of the local power generation equipment is adjusted. gen , to balance the power supply and demand. At the same time, the regional distribution sub-system will continuously attempt to restore communication with the cross-regional intelligent distribution network collaborative scheduling platform. Once the communication is restored, the local operation data will be immediately uploaded to the platform, and subsequent scheduling operations will be carried out according to the instructions issued by the platform.

[0039] Further, the data acquisition and transmission module also has the ability to finely collect distributed energy access data. Distributed energies such as solar energy and wind energy are intermittent and volatile, and their access has an important impact on the operation of the distribution network. Therefore, the data acquisition and transmission module not only collects the total power generation E of distributed energy, but also collects the real-time changes in its power generation power and detailed information on the operation status of the power generation equipment. For solar power generation, parameters such as the temperature T of the photovoltaic panel and the light intensity I that affect the power generation efficiency are collected. For wind power generation, data such as the wind speed v, wind direction d, and rotational speed n of the wind turbine are collected. By collecting and analyzing these fine data, the power generation characteristics of distributed energy can be more accurately grasped, providing more comprehensive information for the scheduling decision-making module, which helps to optimize the access and scheduling of distributed energy in the cross-regional distribution network and improve the energy utilization efficiency and the stability of the power grid.

[0040] Further, the cross-regional intelligent distribution network collaborative scheduling platform also has a user demand response management function. The platform collects information on users' electricity consumption behavior habits, electricity consumption preferences, and adjustable electricity consumption time periods through information interaction with users, and establishes a user demand response model. This model is based on the historical electricity consumption data D of users h and the feedback information of users on different incentive measures. By analyzing historical data, the electricity consumption response laws of users under different electricity prices and different time periods are mined. According to different electricity price policies and the load conditions of the power grid, electricity consumption incentive signals are sent to users. During the peak load period t of the power grid peak , in order to guide users to reduce unnecessary electricity consumption, the platform raises the electricity price P peak , and the incentive signal sent to users at this time can be to inform users that it is currently the peak period and the electricity price is high, and encourage users to reasonably adjust the usage time of electrical equipment. During the low load period t valley , the electricity price P is reduced valley, and send signals to users to guide them to increase electricity consumption. The user demand response model is continuously optimized to improve the prediction accuracy of user behavior. In the model, a user response coefficient r is introduced, which reflects the sensitivity of users to electricity price changes and is determined through statistical analysis of a large amount of user data. Different types of users have different user response coefficients. By adjusting the incentive strategy, a better demand response effect can be achieved. At the same time, the platform monitors and evaluates the response behavior of users in real time, and further optimizes the model and incentive strategy according to the actual response of users. Through this user demand response management function, the energy distribution and scheduling of the cross-regional distribution network can be further optimized, while meeting the electricity consumption needs of users, improving the operation efficiency and resource utilization rate of the power grid, and realizing the positive interaction and coordinated development between the power grid and users.

[0041] On the other hand, a method for dynamic energy scheduling management of an intelligent distribution network is characterized in that the specific steps of the method are as follows:

[0042] S1. Data storage: Store the real-time data and historical scheduling data from each regional distribution sub-system. The real-time data includes the voltage V, current I, power P, and load L data of each regional distribution network, the power generation data E of distributed energy, and the electricity consumption demand data D of users. The historical scheduling data includes the energy distribution plan and scheduling execution result information in different past time periods.

[0043] S2. Data analysis: Use big data analysis algorithms to deeply mine and analyze the collected data, and use the Kalman filtering algorithm to denoise the real-time data. The formula is:

[0044]

[0045] P k|k-1 = AP k-1|k-1 A T + Q

[0046] K k = P k|k-1 H T (HP k|k-1 H T + R) -1

[0047]

[0048] P k|k = (I - K k H)P k|k-1

[0049] Where, is the state prediction value at time k based on the information at time k - 1, A is the state transition matrix, is the state estimate at time k-1, B is the control input matrix, and u k-1 is the control input at time k-1, and P k|k-1 is the prediction error covariance at time k based on the information at time k-1, Q is the process noise covariance, and K k is the Kalman gain, H is the observation matrix, and Z k is the observation at time k, R is the observation noise covariance, and is the state estimate at time k, and P k|k is the estimation error covariance at time k. By analyzing historical data and denoised real-time data, valuable information such as load prediction and energy distribution is extracted. For load prediction, an adaptive-weight neural network algorithm is used. The input layer nodes correspond to different influencing factors, time t, weather W, and historical load L h , and the hidden layer nodes perform non-linear transformation through the activation function , and the output layer is the predicted load value The network weights w and biases are adaptively adjusted by minimizing the mean square error between the predicted value and the actual value , where N is the number of samples L n is the actual load value, is the predicted load value, and the initial values of the weights are generated by random numbers and then updated according to the backpropagation algorithm during the training process;

[0050] S3. Scheduling decision: According to the results of the data analysis module, a multi-objective optimization scheduling algorithm is used to generate the optimal scheduling plan. This algorithm takes resource optimization configuration, operation efficiency improvement, and reliability enhancement as multiple objectives and constructs the objective function:

[0051] F = ω1×C r + ω2×C e + ω3×C rli

[0052] where C r represents the resource allocation cost, C e represents the operation efficiency index, C rli represents the reliability index, and ω1, ω2, and ω3 are the weights corresponding to the respective objectives. During the solution process, the energy supply and demand balance of each region, the transmission capacity of the power grid, and the access constraint conditions of distributed energy are considered, and the solution is carried out through the particle swarm algorithm. In the particle swarm algorithm, the position of the particle represents a scheduling plan, the velocity determines the moving direction and distance of the particle, and the position update formula of the particle is:

[0053] x id (t + 1) = x id (t) + v id (t + 1)

[0054] v id v(t + 1) = ωv id (t) + c1r1(t)(p id -x id (t)) + c2r2(t)(g d -x id (t))

[0055] Wherein, x id (t) is the position of the i-th particle in the d-th dimensional space, v i d(t) is its velocity, ω is the inertia weight, c1 and c2 are learning factors, r1(t) and r2(t) are random numbers between [0, 1], p id is the historical best position of the particle itself, g d is the global best position. The inertia weight ω is dynamically adjusted with the number of iterations to balance the global search and local search capabilities. The formula is:

[0056]

[0057] Wherein, ω max and ω min are the maximum and minimum values of the inertia weight respectively, T max is the maximum number of iterations, and t is the current number of iterations;

[0058] S4. Communication interface: Responsible for high-speed and stable data interaction with each regional power distribution subsystem;

[0059] S5. Multiple regional power distribution subsystems: Responsible for the operation management of the power distribution network within their respective regions, including data acquisition and local dispatching functions. Each regional power distribution subsystem communicates with the cross-regional intelligent power distribution network collaborative dispatching platform through the data acquisition and transmission module, uploads the power grid operation data V, I, P, L, E, D of its own region to the platform, and receives the dispatching instructions issued by the platform;

[0060] S6. Data acquisition and transmission: Adopt high-precision sensors and high-speed communication technologies to collect the operation data V, I, P, L of the power distribution network in each region in real time, and transmit these data accurately to the cross-regional intelligent power distribution network collaborative dispatching platform. The data acquisition frequency f can be dynamically adjusted according to actual needs.

[0061] Compared with the prior art, the intelligent power distribution network energy dynamic dispatching management system and method have the following beneficial effects:

[0062] I. By integrating multiple modules including data storage, data analysis, scheduling decision-making, communication interfaces, regional power distribution subsystems, and data collection and transmission, this system realizes the dynamic scheduling management of the energy of the intelligent power distribution network. Such an integrated management system can collect and analyze in real time the voltage, current, power, load data from each regional power distribution network and the power generation data of distributed energy sources, providing comprehensive and accurate information support for scheduling decision-making. At the same time, the multi-objective optimal scheduling algorithm adopted by the system can comprehensively consider multiple objectives such as optimal resource allocation, improved operation efficiency, and enhanced reliability, generating the optimal scheduling plan, thereby improving the energy utilization efficiency and overall operation performance of the power grid.

[0063] II. Through the data analysis module, this system can use big data analysis algorithms to deeply mine and analyze the collected data, and use the Kalman filtering algorithm to denoise the real-time data, improving the accuracy and reliability of the data. At the same time, the system also has the ability to detect and process abnormal data, ensuring the integrity and accuracy of the data. In addition, when the communication with the cross-regional intelligent power distribution network collaborative scheduling platform is interrupted or the platform fails, the regional power distribution subsystem can automatically switch to the local emergency scheduling mode, and perform energy allocation and scheduling according to the pre-set local scheduling strategy and the historical data stored locally, ensuring that the power consumption needs of important users are met. This emergency response ability helps to improve the risk resistance ability of the power grid and ensure the stability and reliability of power supply.

[0064] Other advantages, objectives, and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0066] Figure 1 It is a flow operation diagram of the intelligent power distribution network energy dynamic scheduling management system;

[0067] Figure 2 It is a flow operation diagram of the intelligent power distribution network energy dynamic scheduling management method. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0068] To further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation manners, structures, features, and effects according to the present invention as follows.

[0069] Embodiment 1

[0070] This embodiment describes a certain city divided into multiple regions, each region having an independent power distribution subsystem. The region contains various industrial and commercial users and residential users, and the power consumption load fluctuates greatly. At the same time, there are distributed energy sources such as solar power stations and small wind farms distributed in the region, and their power generation is significantly affected by the weather.

[0071] Each regional power distribution subsystem collects voltage V, current I, power P, load L data, as well as power generation data E of distributed energy sources (including light intensity I and photovoltaic panel temperature T of solar power stations, wind speed v, wind direction d, and fan speed n of small wind farms) and power consumption demand data D of users through high-precision sensors installed at key grid nodes at a frequency of once every 15 minutes (f = 1 / 15 hour -1 ). These data are transmitted to the data acquisition and transmission module through a high-speed communication network. This module uses the distributed hash table technology to calculate the storage location according to the data characteristics and stores the data in different storage nodes in the data storage module. For example, real-time data is stored in chronological order, and historical data is archived according to different scheduling cycles and regional dimensions. The data storage module adopts a distributed storage architecture, and multiple storage nodes are distributed at different geographical locations in the city and are connected through a high-speed network to ensure the reliability and scalability of the data.

[0072] The data analysis module obtains data from the data storage module and uses the Kalman filtering algorithm to denoise the real-time data. Let the state variable X be the grid operation state (such as voltage and current) at time k. The state transition matrix A, control input matrix B, and observation matrix H are determined according to the grid model, and the process noise covariance Q and observation noise covariance R are set according to the sensor accuracy and grid noise characteristics. Taking voltage as an example, the state prediction value X of time k based on the information at time k - 1 k|k-1 The calculation formula is X k|k-1 = AX k-1|k-1 + Bu k-1 , the prediction error covariance P k |k - 1 = AP k-1|k-1 A T + Q, the Kalman gain K k = P k|k-1 H T (HP k|k-1 H T + R) -1 , the state estimate value X of time k k|k = X k|k-1 + K k (Z k - HX k|k-1 ), the estimated error covariance P k|k = (I - K k H)P k|k-1 , where Z k is the observed value at time k (i.e., the voltage value collected by the sensor). Through the denoised real-time data and historical data, an adaptive weight neural network algorithm is used for load forecasting. The input layer nodes include time t, weather W (such as meteorological factors affecting the load like temperature and humidity), and historical load L h , the hidden layer nodes perform non-linear transformation through the activation function , the output layer is the predicted load value L. The network weights w and biases are adaptively adjusted by minimizing the mean square error between the predicted value and the actual value , where N is the number of samples, L n is the actual load value, L n is the predicted load value. The initial values of the weights are randomly generated and updated according to the backpropagation algorithm during the training process.

[0073] The scheduling decision module generates the optimal scheduling plan according to the data analysis results, using the multi-objective optimization scheduling algorithm. The objective function F = ω1×C r + ω2×C e + ω3×C rli , where ω1, ω2, and ω3 are the weights of the resource allocation cost C r , the operation efficiency index C e , and the reliability index C rli , respectively, which are set according to the focus of the urban power grid operation (such as paying more attention to reliability during peak hours, and the weights can be adjusted accordingly). During the solution process, the energy supply-demand balance in each region (i.e., the total power generation in the region is equal to the total load demand), the power grid transmission capacity (the line transmission power P ij meets determined by the line physical parameters and safety standards), and the distributed energy access constraints (such as the solar power generation power is limited by the light intensity and temperature) are considered. The particle swarm algorithm is used to solve the problem. The particle position represents a scheduling plan, and the velocity determines the moving direction and distance of the particle. The particle position update formula is x id (t + 1) = x id (t) + v id (t + 1), v id (t + 1) = ωv id (t) + c1r1(t)(p id - x id (t)) + c2r2(t)(g d - x id(t)), where i is the particle number, d is the dimension, t is the iteration number, ω is the inertia weight, (ω max and ω min are the maximum and minimum values of the inertia weight respectively, T m ax is the maximum number of iterations), c1 and c2 are learning factors, r1(t) and r2(t) are random numbers between [0, 1], p i d is the historical best position of the particle itself, g d is the global best position. For example, during the peak summer electricity consumption period, by adjusting the output of distributed energy and the power transmission between regions, the load demand is met, while the cost is reduced, and the operation efficiency and reliability are improved.

[0074] The communication interface module uses encrypted communication technology (generating the public key K pub and the private key K pri by the RSA algorithm to encrypt a small amount of key data, and using the symmetric key K sym by the AES algorithm to encrypt a large amount of data, and the symmetric key is encrypted and transmitted by the RSA algorithm) to perform high-speed and stable data interaction with each regional distribution sub-system. The scheduling instructions generated by the scheduling decision module are sent to each regional distribution sub-system through the communication interface module, and the regional distribution sub-system executes the scheduling instructions to adjust the output power of the power generation equipment and the load distribution, realizing the optimal operation of the power grid.

[0075] The above are only the preferred embodiments of the present invention, and do not impose any form of limitation on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the equivalent embodiments with equivalent changes within the scope of the technical solution of the present invention. However, as long as it does not depart from the content of the technical solution of the present invention, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. Intelligent distribution network energy dynamic dispatching and management system, characterized in that: The system includes a data storage module, a data analysis module, a dispatch decision module and a communication interface module as well as multiple regional distribution subsystems and data acquisition and transmission modules: The data storage module stores real-time data and historical dispatch data from distribution subsystems in each region. The real-time data includes voltage V, current I, power P, load L data of distribution networks in each region, power generation data E of distributed energy, and power demand data D of users. The historical dispatch data includes energy allocation plans and dispatch execution result information in different time periods in the past. The data analysis module: uses a large-scale data analysis algorithm to perform in-depth mining and analysis on the collected data, and uses a Kalman filter algorithm to perform denoising on real-time data. The formula is: P k|k-1 =AP k-1|k-1 From T +Q K k =P k|k-1 H T (HP k|k-1 H T +R) -1 P k|k =(I-K k H)P k|k-1 in, is the state prediction value at time k based on the information at time k-1, A is the state transfer matrix, is the state estimate at time k-1, B is the control input matrix, u k-1 is the control input at time k-1, P k|k-1 is the prediction error covariance at time k based on information at time k-1, Q is the process noise covariance, K k is the Kalman gain, H is the observation matrix, Z k is the observed value at time k, R is the observation noise covariance, is the estimated value of the state at time k, P k|k is the estimated error covariance at time k. By analyzing historical data and denoised real-time data, valuable information such as load forecasting and energy distribution is extracted. For load forecasting, an adaptive weighted neural network algorithm is used, and its input layer nodes correspond to different influencing factors, such as time t, weather W, and historical load L. h , the nodes in the acute hidden layer are activated by the function Perform nonlinear transformation, and the output layer is the predicted load value The network weights w and bias are calculated by minimizing the mean square error between the predicted value and the actual value. Perform adaptive adjustment, where N is the number of samples L n is the actual load value, is the predicted load value. The initial value of the weight is generated by random numbers and then updated according to the back-propagation algorithm during the training process; The scheduling decision module: Based on the results of the data analysis module, the multi-objective optimization scheduling algorithm is used to generate the optimal scheduling plan. The algorithm takes resource optimization configuration, operation efficiency improvement, and reliability enhancement as multiple objectives to construct the objective function: F=ω1×C r +ω2×C e +ω3×C rli Among them, C r represents the resource allocation cost, C e represents the operating efficiency index, C rli Represents the reliability index, ω1, ω2, and ω3 are the weights of the corresponding targets. In the solution process, the energy supply and demand balance of each region, the transmission capacity of the power grid, and the access constraints of distributed energy are considered, and the particle swarm algorithm is used to solve it. In the particle swarm algorithm, the position of the particle represents a scheduling scheme, and the speed determines the moving direction and distance of the particle. The particle position update formula is: x id (t+1)=x id (t)+v id (t+1) v id (t+1)=ωv id (t)+c1r1(t)(p id -x id (t))+c2r2(t)(g d -x id (t)) Among them, x id (t) is the position of the ith particle in the d-dimensional space, v i d(t) is its velocity, ω is the inertia weight, c1 and c2 are learning factors, r1(t) and r2(t) are random numbers between [0, 1], and p id is the particle’s own historical optimal position, g d is the global optimal position, and the inertia weight ω is dynamically adjusted with the number of iterations to balance the global search and local search capabilities. The formula is: Among them, ω max and ω min are the maximum and minimum values ​​of the inertia weight, T max is the maximum number of iterations, t is the current number of iterations; The communication interface module is responsible for high-speed and stable data interaction with the distribution subsystems in each area; The multiple regional distribution subsystems are responsible for the operation and management of the distribution network in their respective regions, including data collection and local dispatching functions. Each regional distribution subsystem communicates with the cross-regional intelligent distribution network collaborative dispatching platform through the data collection and transmission module, uploads the power grid operation data V, I, P, L, E, D of the region to the platform, and receives the dispatching instructions issued by the platform; The data acquisition and transmission module uses high-precision sensors and high-speed communication technology to collect the operating data V, I, P, and L of the distribution networks in each region in real time, and accurately transmits these data to the cross-regional intelligent distribution network collaborative dispatching platform. The data acquisition frequency f can be dynamically adjusted according to actual needs.

2. The smart distribution network energy dynamic dispatching and management system according to claim 1 is characterized in that: The data storage module adopts a distributed storage architecture. The data storage module consists of multiple storage nodes. These storage nodes are distributed in different geographical locations and connected through high-speed networks. Distributed hash table technology is used to store and manage data. Each storage node is responsible for storing a part of the data. When new data needs to be stored, the storage location is calculated by the hash function according to the characteristics of the data, and the data is stored in the corresponding storage node. The hash function is designed as follows: Among them, key is the identifier of the data, a i is the coefficient, n is the degree of the polynomial, and m is the number of storage nodes. During data storage, different types of data are classified and stored. Real-time data is stored in chronological order and stored and recorded at regular intervals. Historical data is classified and archived according to different scheduling cycles and regional dimensions.

3. The smart distribution network energy dynamic dispatching and management system according to claim 1 is characterized in that: The data analysis module also has the ability to detect and process abnormal data. During the data collection process, abnormal data may appear due to sensor failure and communication interference. The data analysis module uses a statistical method to detect abnormal data. First, the mean μ and standard deviation σ of the data are calculated: Among them, x i is a data sample, N is the number of samples, and a threshold k is set. When a data point x satisfies |x-μ|>kσ, it is judged as abnormal data. For abnormal data, the interpolation method is used to process it. According to the values ​​and time intervals of adjacent normal data points, the linear interpolation formula is used: Among them, (x1, y1) and (x2, y2) are adjacent normal data points, x is the position of the abnormal data point, and y is the substitute value obtained by interpolation. When processing abnormal values ​​of different types of data, the above unified detection and processing method is followed.

4. The smart distribution network energy dynamic dispatching and management system according to claim 1 is characterized in that: The scheduling decision module also considers the safety constraints of the power grid when generating the scheduling plan. The safe operation of the power grid is an important prerequisite for scheduling. Therefore, safety constraints are added to the multi-objective optimization scheduling algorithm. The node voltage constraints are: In min ≤V i ≤V max Among them, V i is the voltage at node i, V min and V max They are the lower and upper limits of the node voltage, respectively. These two values ​​are determined according to the design standards and operation requirements of the power grid; Line transmission power constraints: Among them, P ij is the transmission power of line ij, is the maximum transmission power of line ij, which is determined by the physical parameters and safety standards of the line. It also takes into account the short-circuit current constraint of the system. Excessive short-circuit current will cause damage to the power grid equipment. Therefore, it is necessary to limit the short-circuit current within a certain range. Suppose the short-circuit current is I sc , and its constraints are Determined according to the tolerance capacity of the power grid equipment, these safety constraints are used as limiting conditions when solving the multi-objective optimization scheduling algorithm.

5. The smart distribution network energy dynamic dispatching and management system according to claim 1 is characterized in that: The communication interface module uses encryption communication technology. During the data interaction process, the communication interface module encrypts the transmitted data and generates a public key K through the RSA algorithm by combining an asymmetric encryption algorithm with a symmetric encryption algorithm. pub and private key K pri The public key is used to encrypt data, and the private key is used to decrypt. Before data transmission, the sender uses the receiver's public key K pab Encrypt the data and then send it out. After receiving the data, the receiver uses his own private key K pri To improve encryption efficiency, the AES algorithm is used to encrypt large amounts of data using a symmetric key K sym Encrypt and decrypt data, the symmetric key K sym The data is encrypted using the RSA algorithm.

6. The smart distribution network energy dynamic dispatching and management system according to claim 1, characterized in that: The regional distribution subsystem has a local emergency dispatch function. When the communication with the cross-regional intelligent distribution network collaborative dispatching platform is interrupted or the platform fails, the regional distribution subsystem can automatically switch to the local emergency dispatch mode. In the local emergency dispatch mode, the regional distribution subsystem performs energy distribution and dispatch according to the pre-set local dispatch strategy and the historical data stored locally. The local dispatch strategy is formulated based on the analysis of the historical operation data and load characteristics of the regional power grid, giving priority to the power demand of important users. The determination of important users is based on the comprehensive judgment of the industry attributes of the users and the stability factors of the power load. For the load demand L of important users, imp , which is given priority in the dispatching process, and the output power of the power generation equipment is adjusted according to the load forecast. The load forecast uses the local historical load data L h , combined with local weather data W and time factor t, through a simple linear regression model L pred =a0+a1t+a2W+a3L h Forecasting is performed, where a0, a1, a2, and a3 are coefficients obtained by fitting historical data. According to the forecast results, the output power P of the local power generation equipment is adjusted. gen In order to balance the supply and demand of electricity, the regional distribution subsystem will continue to try to restore communication with the cross-regional smart distribution network collaborative dispatching platform. Once the communication is restored, the local operation data will be uploaded to the platform immediately, and subsequent dispatching operations will be carried out according to the instructions issued by the platform.

7. The smart distribution network energy dynamic dispatching and management system according to claim 1 is characterized in that: The data acquisition and transmission module also has the ability to collect distributed energy access data in a refined manner. Distributed energy such as solar energy and wind energy are intermittent and volatile, and their access has an important impact on the operation of the distribution network. Therefore, the data acquisition and transmission module not only collects the total power generation E of distributed energy, but also collects the real-time changes in its power generation power and detailed information on the operating status of the power generation equipment. For solar power generation, the temperature T of the photovoltaic panel and the light intensity I, which are parameters affecting the power generation efficiency, are collected. For wind power generation, the wind speed v, wind direction d, and wind turbine speed n data are collected.

8. The smart distribution network energy dynamic dispatching and management system according to claim 1 is characterized in that: The cross-regional intelligent distribution network collaborative dispatching platform also has a user demand response management function. The platform collects users' electricity consumption behavior habits, electricity consumption preferences and adjustable electricity consumption time information through information interaction with users, and establishes a user demand response model. The model is based on the user's historical electricity consumption data D h As well as the feedback information of users on different incentive measures, through the analysis of historical data, the user's power consumption response pattern under different electricity prices and different time periods is explored, and according to different electricity price policies and grid load conditions, power consumption incentive signals are sent to users. peak In order to guide users to reduce unnecessary electricity consumption, the platform increases the electricity price P peak At this time, the incentive signal sent to the user can be to inform the user that the current peak period is high and the electricity price is high, encouraging the user to reasonably adjust the use time of the electrical equipment and use it during the low load period t valley , reduce the electricity price P valley , and send signals to users to guide them to increase electricity consumption. The user demand response model improves the prediction accuracy of user behavior through continuous optimization. In the model, the user response coefficient r is introduced. This coefficient reflects the user's sensitivity to changes in electricity prices. It is determined through statistical analysis of a large amount of user data that different types of users have different user response coefficients. By adjusting the incentive strategy, the platform monitors and evaluates the user's response behavior in real time, and further optimizes the model and incentive strategy according to the user's actual response.

9. A method for dynamic dispatching and managing energy in a smart distribution network, characterized in that: The specific steps of this method are: S1. Data storage: Stores real-time data and historical dispatch data from distribution subsystems in various regions. Real-time data includes voltage V, current I, power P, load L data of distribution networks in various regions, distributed energy generation data E, and user power demand data D. Historical dispatch data includes energy allocation plans and dispatch execution result information in different time periods in the past. S2. Data analysis: Use big data analysis algorithms to conduct in-depth mining and analysis of the collected data, and use the Kalman filter algorithm to denoise the real-time data. The formula is: P k|k-1 =AP k-1|k-1 From T +Q K k =P k|k-1 H T (HP k|k-1 H T +R) -1 P k|k =(I-K k H)P k|k-1 in, is the state prediction value at time k based on the information at time k-1, A is the state transfer matrix, is the state estimate at time k-1, B is the control input matrix, u k-1 is the control input at time k-1, P k|k-1 is the prediction error covariance at time k based on information at time k-1, Q is the process noise covariance, K k is the Kalman gain, H is the observation matrix, Z k is the observed value at time k, R is the observation noise covariance, is the estimated value of the state at time k, P k|k is the estimated error covariance at time k. By analyzing historical data and denoised real-time data, valuable information such as load forecasting and energy distribution is extracted. For load forecasting, an adaptive weighted neural network algorithm is used, and its input layer nodes correspond to different influencing factors, such as time t, weather W, and historical load L. h , the nodes in the acute hidden layer are activated by the function Perform nonlinear transformation, and the output layer is the predicted load value The network weights w and bias are calculated by minimizing the mean square error between the predicted value and the actual value. Perform adaptive adjustment, where N is the number of samples L n is the actual load value, is the predicted load value. The initial value of the weight is generated by random numbers and then updated according to the back-propagation algorithm during the training process; S3, Scheduling decision: Based on the results of the data analysis module, the multi-objective optimization scheduling algorithm is used to generate the optimal scheduling plan. The algorithm takes resource optimization configuration, operation efficiency improvement, and reliability enhancement as multiple objectives to construct the objective function: F=ω1×C r +ω2×C e +ω3×C rli Among them, C r represents the resource allocation cost, C e represents the operating efficiency index, C rli Represents the reliability index, ω1, ω2, and ω3 are the weights of the corresponding targets. In the solution process, the energy supply and demand balance of each region, the transmission capacity of the power grid, and the access constraints of distributed energy are considered, and the particle swarm algorithm is used to solve it. In the particle swarm algorithm, the position of the particle represents a scheduling scheme, and the speed determines the moving direction and distance of the particle. The particle position update formula is: x id (t+1)=x id (t)+v id (t+1) v id (t+1)=ωv id (t)+c1r1(t)(p id -x id (t))+c2r2(t)(g d -x id (t)) Among them, x id (t) is the position of the ith particle in the d-dimensional space, v i d(t) is its velocity, ω is the inertia weight, c1 and c2 are learning factors, r1(t) and r2(t) are random numbers between [0, 1], and p id is the particle’s own historical optimal position, g d is the global optimal position, and the inertia weight ω is dynamically adjusted with the number of iterations to balance the global search and local search capabilities. The formula is: Among them, ω max and ω min are the maximum and minimum values ​​of the inertia weight, T max is the maximum number of iterations, t is the current number of iterations; S4, communication interface: responsible for high-speed and stable data interaction with the distribution subsystems in each region; S5, multiple regional distribution subsystems: responsible for the operation and management of the distribution network in their respective regions, including data collection and local dispatching functions. Each regional distribution subsystem communicates with the cross-regional intelligent distribution network collaborative dispatching platform through the data collection and transmission module, uploads the power grid operation data V, I, P, L, E, D of the region to the platform, and receives the dispatching instructions issued by the platform; S6. Data collection and transmission: Use high-precision sensors and high-speed communication technology to collect the operating data V, I, P, L of each regional distribution network in real time, and transmit these data accurately to the cross-regional intelligent distribution network collaborative dispatching platform. The data collection frequency f can be dynamically adjusted according to actual needs.

Images