Power network distributed intelligent power distribution method and system

Abstract

The invention discloses a distributed intelligent power distribution method and system for a power network, and belongs to the technical field of power distribution control, and the method comprises the steps: collecting multi-dimensional electrical parameters of a power generation node in real time, and carrying out the dynamic power monitoring at a feeder terminal of a transformer substation through an edge calculation node; executing line loss calculation and load prediction in the power distribution unit, generating a demand feedback instruction to a feeder terminal node, and triggering an adaptive power distribution strategy; and on the basis of a digital twin system of a regional dispatching center, credibility verification is carried out on operation decisions of all edge nodes, and global optimization compensation control is executed. The method has the remarkable advantages in the aspects of optimizing power distribution, reducing line loss, improving system stability and safety and achieving efficient energy utilization, and intelligent management of a power network becomes possible.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution control, and particularly relates to a power network distributed intelligent power distribution method and system. BACKGROUND

[0002] Modern power systems are gradually transforming from traditional centralized generation mode to diversified structure with distributed generation as the core. This transformation has given rise to the application of a large number of emerging technologies, such as edge computing, Internet of Things (IoT), blockchain, and deep learning. The combination of these technologies supports efficient management of distributed energy resources and improves the reliability and economy of power supply. At the same time, the stability of the power system and the optimization of the function need to be solved, especially in dealing with power demand fluctuations and equipment failures, the traditional static control method has been shown to be inadequate.

[0003] Although existing intelligent power distribution and monitoring systems have improved the efficiency of power grid operation to some extent, there are still many deficiencies. On the one hand, existing systems often rely on centralized data processing mode, resulting in insufficient real-time response capability, and unable to effectively monitor and analyze dynamic electrical characteristics. On the other hand, traditional line loss calculation methods are mostly static models, which cannot adapt to the actual power flow demand under complex environmental and dynamic load conditions. More seriously, in the monitoring of safe operation of equipment, due to the lack of flexible feedback mechanism, it is easy to cause blind operation of the system under abnormal conditions, increasing the risk of failure. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] Therefore, the technical problem solved by the present application is to solve the technical problem of real-time monitoring of electrical parameters, dynamic calculation of line loss and optimization of load distribution in a distributed power network.

[0006] To solve the above technical problems, the present application provides the following technical solutions: a power network distributed intelligent power distribution method, comprising,

[0007] Real-time acquisition of multi-dimensional electrical parameters of power generation nodes, dynamic power monitoring at the transformer substation feeder terminal through edge computing nodes; line loss calculation and load prediction are performed at the power distribution unit, and demand feedback instructions are generated to the feeder terminal node to trigger adaptive power distribution strategy; relying on the digital twin system of the regional dispatch center, the credibility of the operation decisions of all edge nodes is verified, and global optimization compensation control is performed.

[0008] As a preferred scheme of the power network distributed intelligent power distribution method, the multi-dimensional electrical parameters include three-phase voltage, current harmonic components and real-time power factor collected through multi-dimensional dynamic characteristics.

[0009] As a preferred scheme of the power network distributed intelligent power distribution method, the edge computing node comprises a fog computing architecture, is deployed at all feeder terminals and power distribution units, and is configured with an embedded load mode recognition algorithm.

[0010] The nodes realize safe data sharing through a blockchain ledger and build a distributed consensus decision mechanism.

[0011] As a preferred scheme of the power network distributed intelligent power distribution method, the dynamic power monitoring comprises monitoring instantaneous active power received by a substation from a multi-source power generation end through an edge intelligent terminal.

[0012] As a preferred scheme of the power network distributed intelligent power distribution method, the line loss calculation and load prediction comprise establishing a dynamic line loss model based on input power of dynamic power monitoring.

[0013] The load time series data uploaded by the aggregated user side smart meters is aggregated.

[0014] A demand prediction model is established by using a deep residual network to fuse space-time features.

[0015] The demand index of the power distribution unit is fed back.

[0016] As a preferred scheme of the power network distributed intelligent power distribution method, the adaptive power distribution strategy comprises constructing a power fluctuation spectrum analysis model.

[0017] A multi-objective optimization function is established.

[0018] The optimal distribution scheme is solved by using the Lagrange multiplier method.

[0019] As a preferred scheme of the power network distributed intelligent power distribution method, the global optimization compensation control comprises,

[0020] A stability evaluation model based on a deep Q network is constructed, and a device safe operation interval is output.

[0021] When the optimal distribution scheme belongs to the device safe operation interval, the bidirectional power flow control of the hybrid energy storage system is activated.

[0022] A source-storage-load dynamic balance equation is established for compensation.

[0023] Another object of the present application is to provide a power network distributed intelligent power distribution system.

[0024] To solve the above technical problems, the application provides the following technical solutions: a power network distributed intelligent power distribution system, comprising a data sensing module, an edge decision module, a power distribution module and a monitoring optimization module.

[0025] The data sensing module collects multi-dimensional electrical parameters of a power generation node in real time.

[0026] The edge decision module implements dynamic power monitoring at a feeder terminal of a substation through an edge computing node.

[0027] The power distribution module performs line loss calculation and load prediction at a power distribution unit, generates a demand feedback instruction to the feeder terminal node, and triggers an adaptive power distribution strategy.

[0028] The monitoring optimization module relies on a digital twin system of a regional dispatch center to verify the credibility of the operation decisions of all edge nodes and perform global optimization compensation control.

[0029] The application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the power network distributed intelligent power distribution method when executing the computer program.

[0030] The application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to implement the steps of the power network distributed intelligent power distribution method.

[0031] The application has the following beneficial effects: The application realizes comprehensive state monitoring of power generation nodes by collecting multi-dimensional electrical parameters of power generation nodes in real time and implementing dynamic power monitoring, effectively identifies abnormal conditions, and improves the operation safety and reliability of the power distribution system.

[0032] After the power distribution unit performs line loss calculation and load prediction, the application generates a dynamic line loss model and a load instruction, so that the power system can quickly respond to load changes and optimize the power distribution strategy, reduce energy loss, and improve the accuracy of line loss estimation by combining environmental factors.

[0033] The application aggregates user-side smart meter data, establishes a demand prediction model using deep learning, realizes intelligent scheduling and optimization, verifies the credibility of edge node operation decisions relying on a digital twin system, ensures the effectiveness of the optimization scheme, improves the intelligentization and adaptive ability of the system, and provides additional protection. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and all other drawings obtained by those skilled in the art without creative effort based on these drawings should also fall within the protection scope of the present application.

[0035] Figure 1 A general flowchart of a power network distributed intelligent power distribution method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings. Obviously, the described embodiments are only some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should also fall within the protection scope of the present application.

[0037] Embodiment 1, refer to Figure 1 For an embodiment of the present application, the embodiment provides a power network distributed intelligent power distribution method, comprising:

[0038] S1, real-time collection of multi-dimensional electrical parameters of a power generation node, dynamic power monitoring is implemented at a feeder terminal of a substation through an edge computing node;

[0039] Further, multi-dimensional dynamic characteristics of the power generation unit are collected through a broadband synchronous measurement device, including:

[0040] The transient process quantity includes but is not limited to d-axis / q-axis current component, subsynchronous oscillation mode, the steady-state parameter includes but is not limited to total harmonic distortion (THD), three-phase unbalance degree, the operating state quantity includes but is not limited to inverter switching frequency, energy storage SOC;

[0041] In addition, the collection of multi-dimensional electrical parameters includes three-phase voltage, current harmonic component and real-time power factor.

[0042] Further, the edge computing node includes an edge decision cluster composed of intelligent terminals with FPGA acceleration, which is deployed at all feeder terminals and power distribution units and configured with embedded load mode recognition algorithm.

[0043] The nodes realize safe data sharing through a blockchain ledger and build a distributed consensus decision mechanism.

[0044] It should be noted that the edge intelligent terminal calculates the instantaneous active power P(t) received by the substation from the multi-source power generation end.

[0045] S2, performing line loss calculation and load prediction at the distribution unit to generate demand feedback instructions to the feeder terminal node, triggering adaptive distribution strategy;

[0046] Further, in embodiments of the present application, the active power P in (t) is based on the input of dynamic power monitoring in (t)) / P(t)×100.

[0047] In an optional embodiment, the dynamic line loss model is an edge computing node with IP67 protection level deployed at the feeder terminal, which performs:

[0048] An improved Hampel filtering algorithm is used, with window size W=2f s , f s being the sampling frequency.

[0049] Dynamic line loss modeling is performed:

[0050]

[0051] where k is the sampling point index, s and r represent the sending end and the receiving end, respectively.

[0052] η(t) represents the instantaneous line loss rate at time t, i.e., the proportion of power loss in the power transmission process at a specific time t; N is the number of sampling points, Us(k) represents the effective value of the voltage at the sending end s at the kth sampling point; Is(k) represents the effective value of the current at the sending end s at the kth sampling point; cosθ s represents the power factor of the sending end s, θ s is the phase angle between the voltage and the current at the sending end. Ur(k) represents the effective value of the voltage at the receiving end at the kth sampling point; Ir(k) represents the effective value of the current at the receiving end at the kth sampling point; θ r is the phase angle between the voltage and the current at the receiving end.

[0053] In another optional embodiment, IoT sensors are installed at key distribution nodes and feeder terminals to monitor parameters such as current, voltage, and environmental temperature in real time; through an IoT platform, the data collected by each sensor is integrated into a unified database to ensure real-time and completeness of the data; in combination with the real-time collected electrical parameters, a line loss estimation model is dynamically generated, considering dynamic factors such as temperature affecting the change of resistance; a threshold is set, and once abnormal line loss or electrical parameter change is detected, an alarm is sent in real time to facilitate prompt response by relevant personnel; the line loss monitoring results are fed back to the edge computing node to perfect the adaptive distribution strategy, achieving dynamic optimization.

[0054] Then, the load time series data uploaded by the user side smart meter is aggregated;

[0055] A demand forecasting model is established by using a deep residual network to fuse the space-time features:

[0056] D(T+h) = σ(ω ar ARIMA(p,d,q) + ω tcn TCN(X,T)) S(T)

[0057] where ARIMA is a difference integrated moving average model, TCN is a time domain convolution network, and S(T) is a periodic component decomposition result; D(T+h) represents the demand forecast value at future time T+h, and demand usually refers to the maximum average power in a period of time (such as 15 minutes); T is the current time reference point, and h is the future time span of the prediction; σ represents the ReLU activation function, which performs a nonlinear transformation on the weighted combination result of the models in the parentheses, so that it is suitable for the range and characteristics of the prediction target; ω ar represents the weight coefficient allocated to the prediction result of the ARIMA model, and ar is the subscript, representing ARIMA; ARIMA(p,d,q) represents the output of the ARIMA model at time T for the demand forecast value at future time T+h, p is the autoregressive order in the ARIMA model, d is the difference order in the ARIMA model, and q is the moving average order in the ARIMA model; ω tcn represents the weight coefficient allocated to the prediction result of the TCN model, and cn is the subscript, representing TCN; TCN(X,T) represents the output of the TCN model at time T under the condition of input feature X for the demand forecast value at future time T+h, X is the feature data input to the time domain convolution network, which should include the load time series data uploaded by the user side smart meter and its related data, and T represents the current time reference point.

[0058] A feedback demand index of the power distribution unit is generated:

[0059] Q n = D(T+h) / (1-η)

[0060] where Q n represents the feedback constant index generated by the power distribution unit, which takes into account the predicted demand and line loss, and is used to guide the adjustment of power distribution strategy; n represents the node index, and η represents the line loss rate.

[0061] Further, a power fluctuation spectrum analysis model is constructed:

[0062]

[0063] where FPM i represents the power fluctuation spectrum feature of the i-th feeder, and i is the subscript, representing the feeder index; W represents the inputted historical power time series data weather W represents the inputted weather feature vector, containing weather factors (such as temperature, humidity, light intensity, wind speed, etc.) related to power fluctuation.

[0064] A multi-objective optimization function is established;

[0065]

[0066] wherein, K margin is a system safety margin coefficient; TDM i represents the final target demand allocated to the ith object, R represents the total number of feeders participating in optimization in the system, s.t. represents a constraint condition, and Z represents a total load benchmark value or minimum total power supply requirement that the system needs to meet.

[0067] In an embodiment of the present application, the Lagrange multiplier method is used to solve the optimal allocation scheme:

[0068]

[0069] wherein, represents the optimal target demand allocation value of the ith object, λ represents a Lagrange multiplier, and Δ total represents a total deviation amount of the constraint condition, which is usually defined as i.e., the total constraint value minus the sum of all object basic demand predictions; K remaining represents the remaining allocable margin, Δ i represents the ith component of the vector Δ.

[0070] In an optional embodiment, solving the optimal allocation scheme can be to generate a plurality of random target demand allocation schemes to form an initial population; to calculate the fitness of each scheme, which can evaluate the pros and cons of the scheme through indicators such as line loss and power supply reliability; to select the better scheme according to the fitness to participate in the next round of reproduction, thereby forming a new population; to perform crossover on the selected scheme to generate a new scheme, which combines the advantages of different schemes; to perform a small random adjustment on the newly generated scheme to increase the diversity of the population; to repeat the fitness evaluation to the random adjustment until the fitness reaches a preset threshold or reaches a maximum number of iterations; and to select the scheme with the highest fitness as the final target demand allocation scheme.

[0071] In another optional embodiment, solving the optimal allocation scheme can also be to create a reinforcement learning agent that will learn how to make demand allocation decisions in different states; define the actions that the agent can choose, such as allocating different levels of target demand to different feeders; design a reward mechanism that gives positive or negative feedback according to the actual line loss and power supply reliability generated after each allocation, to promote the agent's optimization decision; let the agent interact with the environment in multiple cycles, update its strategy, and adjust the demand allocation strategy according to the rewards obtained; after multiple training, the agent can give the best target demand allocation scheme under the current state.

[0072] S3, the digital twin system relying on the regional dispatching center verifies the credibility of the operation and decision of all edge nodes, and executes global optimization compensation control.

[0073] Further, in the embodiments of the present application, a digital twin virtual mirror system containing 10,000+ nodes is constructed, and the optimization result (such as the optimal allocation scheme) is taken as input to perform credibility verification in the digital twin system, and a stable evaluation model based on deep Q network is output to obtain the device safe operation interval

[0074] When there is an optimal allocation scheme the bidirectional power flow control of the hybrid energy storage system is activated:

[0075]

[0076] where β is the energy storage response coefficient, saturate() is the amplitude limiting function (safe upper limit power), P ess (t) represents the charge / discharge power of the hybrid energy storage system ESS at time t.

[0077] In combination with the bidirectional power flow control of the activated hybrid energy storage system, a source-storage-load dynamic balance equation is established to compensate the real-time power deviation:

[0078] ΣP gen +ΣP ess =ΣP load +ΣP loss

[0079] where ΣP gen represents the sum of the total active power output by all power generation units, ΣP ess represents the sum of the net active power output by all energy storage systems, ΣP load represents the sum of the total active power consumed by all loads (Load), and ΣP loss represents the sum of the total active power loss of the system (line loss, transformer loss, etc.).

[0080] Embodiment 2, as an embodiment of the present application, provides a power network distributed intelligent power distribution system, comprising: a data sensing module, an edge decision module, a power distribution module, and a monitoring optimization module.

[0081] The data sensing module collects multi-dimensional electrical parameters of a power generation node in real time.

[0082] The edge decision module implements dynamic power monitoring at a feeder terminal of a substation through an edge computing node.

[0083] The power distribution module performs line loss calculation and load prediction at a power distribution unit, generates a demand feedback instruction to the feeder terminal node, and triggers an adaptive power distribution strategy.

[0084] The monitoring optimization module relies on a digital twin system of a regional dispatch center to verify the credibility of the operation decisions of all edge nodes and perform global optimization compensation control.

[0085] The embodiment also provides an electronic device suitable for a power network distributed intelligent power distribution method, comprising: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions to realize the power network distributed intelligent power distribution method as proposed in the above embodiment.

[0086] The embodiment also provides a storage medium having a computer program stored thereon, and the program is executed by a processor to realize the power network distributed intelligent power distribution method as proposed in the above embodiment.

[0087] The storage medium proposed in the embodiment and the power network distributed intelligent power distribution method proposed in the above embodiment belong to the same inventive concept, and the technical details not described in detail in the embodiment can be referred to the above embodiment, and the embodiment has the same beneficial effects as the above embodiment.

[0088] From the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized by means of software and necessary general hardware, and of course can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a FLASH memory, a hard disk or an optical disk, etc., including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods of various embodiments of the present application.

[0089] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced, without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A distributed intelligent power distribution method for power networks, characterized in that: include, Real-time acquisition of multi-dimensional electrical parameters of power generation nodes, and dynamic power monitoring at substation feeder terminals via edge computing nodes; In the distribution unit, line loss calculation and load forecasting are performed, and demand feedback instructions are generated to the feeder terminal node to trigger the adaptive distribution strategy. Relying on the digital twin system of the regional dispatch center, the credibility of the computational decisions of all edge nodes is verified, and global optimization and compensation control is executed.

2. The distributed intelligent power distribution method for power networks as described in claim 1, characterized in that: The multi-dimensional electrical parameters include the acquisition of three-phase voltage, current harmonic components, and real-time power factor through multi-dimensional dynamic features.

3. The distributed intelligent power distribution method for power networks as described in claim 2, characterized in that: The edge computing nodes include a fog computing architecture, deployed in all feeder terminals and power distribution units, and configured with an embedded load pattern recognition algorithm. Nodes achieve secure data sharing through a blockchain ledger and build a distributed consensus decision-making mechanism.

4. The distributed intelligent power distribution method for power networks as described in claim 3, characterized in that: The dynamic power monitoring includes calculating the instantaneous active power received by the substation from the multi-source power generation end through an edge intelligent terminal.

5. The distributed intelligent power distribution method for power networks as described in claim 4, characterized in that: The execution of line loss calculation and load prediction includes establishing a dynamic line loss model based on the input power of dynamic power monitoring; Aggregate load time-series data uploaded by smart meters on the user side; A demand prediction model is established by fusing spatiotemporal features using a deep residual network. Generate demand indicators for power distribution unit feedback.

6. The distributed intelligent power distribution method for power networks as described in claim 5, characterized in that: The adaptive power distribution strategy includes constructing a power fluctuation spectrum analysis model; Establish a multi-objective optimization function; The optimal allocation scheme is solved using the Lagrange multiplier method.

7. A distributed intelligent power distribution method for power networks as described in claim 6, characterized in that: The global optimization compensation control includes... Construct a stability assessment model based on deep Q-networks and output the safe operating range of the device; When an optimal allocation scheme exists that falls within the safe operating range of the equipment, activate the bidirectional power flow control of the hybrid energy storage system. A dynamic equilibrium equation for the source-storage-load is established for compensation.

8. A distributed intelligent power distribution system for a power network, employing the distributed intelligent power distribution method for a power network as described in any one of claims 1 to 7, characterized in that, include: Data sensing module, edge decision-making module, power distribution module, and monitoring optimization module; The data sensing module collects multi-dimensional electrical parameters of the power generation nodes in real time. The edge decision module performs dynamic power monitoring at the substation feeder terminal through edge computing nodes; The power distribution module performs line loss calculation and load forecasting in the power distribution unit, generates demand feedback instructions to the feeder terminal node, and triggers the adaptive power distribution strategy. The monitoring and optimization module, relying on the digital twin system of the regional dispatch center, verifies the credibility of the computational decisions of all edge nodes and executes global optimization and compensation control.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the distributed intelligent power distribution method for a power network as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the distributed intelligent power distribution method for a power network as described in any one of claims 1 to 7.

Images