Search

CN-122026473-A - Distributed collaborative optimization operation method and system for traffic infrastructure micro-grid cluster

CN122026473ACN 122026473 ACN122026473 ACN 122026473ACN-122026473-A

Abstract

The invention discloses a distributed collaborative optimization operation method and system of a traffic infrastructure micro-grid cluster, and belongs to the technical field of distributed energy system control. The method aims at the problems of poor energy complementation, lower overall economy, high single-point fault risk existing in a centralized control architecture and the like in an independent operation mode of the existing traffic infrastructure micro-networks, and the method comprises the steps that a plurality of traffic infrastructure micro-networks are organized into clusters, each micro-network calculates optimized operation parameters including power instructions by utilizing a local optimization model based on local operation information and shared information from adjacent micro-networks, and a distributed consensus algorithm is adopted for iterative updating so as to achieve agreement on a global target for minimizing the total operation cost of the clusters, and finally each micro-network adjusts and executes operation strategies according to consensus results. The method is mainly used for improving the energy utilization efficiency, the operation economy and the system reliability of the traffic infrastructure micro-grid cluster and supporting the power grid interaction and sustainable traffic energy management.

Inventors

  • LI KAI
  • QI ZHAOCHEN

Assignees

  • 中交投资咨询(北京)有限公司

Dates

Publication Date
20260512
Application Date
20251219

Claims (10)

  1. 1. The distributed collaborative optimization operation method of the traffic infrastructure micro-grid cluster is characterized by comprising the following steps of: S1, forming a plurality of traffic infrastructure micro-networks into a cluster, wherein each micro-network comprises an energy generation unit, an energy storage unit and a load unit, and the load unit at least comprises an electric automobile charging pile, a traffic signal lamp or a street lamp; S2, each micro-grid calculates the optimized operation parameters of the micro-grid through a local optimization model based on local information and shared information from adjacent micro-grids, wherein the operation optimization parameters comprise power instructions for controlling an energy generation unit, an energy storage unit and a load unit, the local information comprises local energy generation amount, storage state and load demand, and the shared information from the adjacent micro-grids at least comprises marginal cost information representing energy supply and demand states of the adjacent micro-grids and global consistency variables of the adjacent micro-grids in a distributed consensus algorithm; s3, enabling a plurality of micro networks to be iteratively updated based on the shared information through a distributed consensus algorithm, and achieving agreement on a global optimization target, wherein the global optimization target is to minimize the total running cost of the cluster; And S4, updating the operation parameters of each micro-network according to the achieved consensus, and executing the optimized operation.
  2. 2. The distributed collaborative optimization operation method according to claim 1, wherein the iterative updating by the distributed consensus algorithm to agree on a global optimization goal comprises: s31, constructing the global optimization target as a global optimization problem comprising a plurality of local variables of the micro-grid and at least one global coupling constraint, wherein the global coupling constraint is a switching power constraint of a cluster and a superior power grid or a power balance constraint inside the cluster; S32, decomposing the global optimization problem into independent local optimization sub-problems of each micro-grid by adopting an alternate direction multiplication sub-method framework, and coupling the sub-problems by introducing global consistency constraint and dual variables; S33, the local decision variable of each micro-grid is a vector, components of the local decision variable form an alternative scheme of the operation optimization parameter, each micro-grid solves a local optimization sub-problem based on the current dual variable and the global consistency variable in each iteration period k, the local decision variable is updated, the updated local decision variable is exchanged among the micro-grids, and the global consistency variable and the dual variable are updated; S34, repeating the step S33 until a preset convergence condition is met, extracting a corresponding component from the finally determined local decision variable, and issuing the component to each unit for execution as an operation optimization parameter.
  3. 3. The distributed collaborative optimization operation method according to claim 2, wherein the distributed collaborative optimization algorithm adopts an asynchronous communication mode, wherein each micro-network performs local calculation and collaborative update immediately after receiving the collaborative intermediate variable of the adjacent micro-network to reduce the influence of communication delay on the collaborative speed, and the collaborative intermediate variable comprises the updated local decision variable or dual variable.
  4. 4. The distributed collaborative optimization operation method according to claim 1, wherein each micro-grid predicts its short-term energy generation and load demand using a prediction model based on historical data and real-time data when locally calculating the optimization operation parameters, and takes the prediction result as input of the local optimization model to improve optimization accuracy and adaptability.
  5. 5. The distributed co-optimization method of claim 4, wherein the predictive model is implemented by a hybrid deep learning model based on an encoder-decoder architecture; each micro-grid uses a predictive model based on historical data and real-time data to predict its short-term energy generation and load requirements when locally calculating optimal operating parameters, including in particular: The system comprises a prediction model deployed in each micro-grid, wherein the prediction model is input and at least comprises historical time sequence data, real-time monitoring data and traffic associated data, the historical time sequence data comprises historical output and load historical data of a local distributed power supply, the real-time monitoring data comprises current illumination intensity, wind speed and energy storage unit charge state, the traffic associated data comprises real-time traffic flow of an associated road, queuing state of an electric vehicle charging station in a service area and charging reservation information, the prediction model is executed in a rolling mode at a frequency higher than an optimal calculation period, and the latest acquired data is utilized to output a predicted sequence of energy generation and load demand of one or more future optimal periods; The workflow of the prediction model comprises three core steps of multi-mode data coding, attention mechanism fusion and sequence prediction decoding, wherein in the multi-mode data coding stage, the prediction model extracts local space-time modes in historical energy and load data through parallel one-dimensional convolutional neural network branches, the time-sequence dependency relationship is coded through long-short-term memory network branches, and weather type and date type static features are processed through full-connection neural network branches; And injecting a prediction sequence output by the prediction model into the local optimization model as basic input, wherein the local optimization model adopts a robust optimization method, expresses an error boundary of the prediction sequence as an uncertainty set, and constructs an optimization problem aiming at minimizing the running cost under the worst condition to solve, so as to generate a robust running optimization parameter immune to the prediction error within a certain range.
  6. 6. The distributed collaborative optimization operation method according to claim 5, wherein the calculating the optimization operation parameters through the local optimization model specifically comprises: Constructing a rolling optimization model with multiple time scales, and obtaining an objective function minimization solution in each optimization period by utilizing the rolling optimization model, wherein the objective function is represented as minimum: alpha multiplied by C energy + β×C degradation + γ×C penalty , wherein C energy is energy cost comprising electricity purchasing cost from an upper-level power grid, electricity selling income to the power grid and power exchanging settlement cost with an adjacent micro-grid, C degradation is life damage cost of a battery energy storage system quantized based on a rain flow counting method principle due to charge and discharge circulation, C penalty is penalty cost comprising power supply reliability constraint violation for key traffic load and electric automobile charging service satisfaction reduction, and alpha, beta and gamma are weight coefficients for adjusting priority among economy, equipment life and power supply reliability under different operation scenes; The decision variables of the rolling optimization model comprise charging and discharging power of an energy storage unit, exchange power with a cluster network, switching states of schedulable loads and real-time output power of an electric vehicle charging pile; The constraint conditions of the optimization model at least comprise system power balance constraint, energy storage unit charge state dynamic constraint considering a life model, distributed energy output upper and lower limit constraint, uninterruptible power supply constraint of key traffic load and total power and minimum service power constraint of an electric vehicle charging pile.
  7. 7. The distributed collaborative optimization operation method according to claim 1, wherein the distributed consensus algorithm comprises a fault tolerance mechanism, when a certain micro-grid fault or communication interruption is detected, the cluster automatically adjusts the consensus process, bypasses the fault micro-grid to continue operation, and maintains global consistency through redundant information exchange.
  8. 8. The distributed collaborative optimization operation method according to claim 1, wherein the distributed consensus algorithm employs an event trigger mechanism wherein each micro-net sends information to neighboring micro-nets only when a local state change exceeds a preset threshold to reduce communication overhead and computing resource consumption.
  9. 9. The distributed collaborative optimization operation method according to claim 1, wherein the local optimization model of each micro-grid comprises a multi-objective optimization function, and energy cost, load priority and environmental impact are considered, and optimization balance under different scenes is achieved through weight adjustment.
  10. 10. A distributed co-optimized operation system for a traffic infrastructure microgrid cluster for implementing the method of claim 1, comprising: Each micro-grid unit comprises an energy generation module, an energy storage module, a load module and a local controller, wherein the local controller is configured to execute local optimization calculation and a distributed consensus algorithm; the communication network is connected with a plurality of micro-network units and used for supporting point-to-point or broadcast communication and exchanging optimization information and consensus data; a global coordinator configured to initialize cluster parameters, monitor global status, and intervene in adjustments when abnormal, but not exercise centralized control.

Description

Distributed collaborative optimization operation method and system for traffic infrastructure micro-grid cluster Technical Field The invention relates to the technical field of distributed energy system control. More particularly, the invention relates to a distributed collaborative optimization operation method and system of a traffic infrastructure micro-grid cluster. Background In the field of traffic infrastructure, micro-grid technology is increasingly used, which is generally composed of distributed energy sources, energy storage units and traffic-related loads. The existing operation modes are mainly divided into two types, namely independent operation of the micro-grid and centralized cluster control. Under the independent operation mode, each micro-grid is optimized only according to the supply and demand state of the micro-grid, and energy complementation and cooperative scheduling in the cluster range are difficult to realize. Due to the lack of an effective coordination mechanism, the phenomenon of coexistence of local energy surplus and shortage easily occurs, so that the overall operation economy is poor, and the cluster aggregation effect cannot be fully utilized to participate in the power grid interaction. Cluster mode with centralized control, while enabling global optimization, is highly dependent on one central control unit. This architecture presents a single point of failure risk, which can lead to a breakdown of the overall system operation once the central node fails. Meanwhile, as the cluster scale expands, the amount of data to be processed by the central controller increases sharply, and extremely high requirements are put on communication bandwidth and computing capacity, so that the system expandability is limited. In addition, centralized control requires all micro-networks to upload detailed operational data to a central node, involving data privacy and security risks. In practical engineering, attempts have also been made to solve the above problems by a distributed method, but many difficulties are faced. On one hand, how to achieve coordination and consistency of global optimization targets is a challenge on the premise of guaranteeing information privacy and decision autonomy of each micro-grid. On the other hand, there is a considerable difficulty in designing a distributed algorithm that can still maintain robustness under non-ideal conditions such as communication delay, data packet loss, and even partial node failure. In addition, loads in the traffic infrastructure microgrid, such as electric car charging piles, have significant randomness and volatility, further increasing the complexity of distributed collaborative optimization. Disclosure of Invention It is an object of the present invention to solve at least the above problems and to provide at least the advantages to be described later. To achieve these objects and other advantages and in accordance with the purpose of the invention, there is provided a distributed collaborative optimization operation method of a traffic infrastructure micro-grid cluster, comprising the steps of: S1, forming a plurality of traffic infrastructure micro-networks into a cluster, wherein each micro-network comprises an energy generation unit, an energy storage unit and a load unit, and the load unit at least comprises an electric automobile charging pile, a traffic signal lamp or a street lamp; S2, each micro-grid calculates the optimized operation parameters of the micro-grid through a local optimization model based on local information and shared information from adjacent micro-grids, wherein the operation optimization parameters comprise power instructions for controlling an energy generation unit, an energy storage unit and a load unit, the local information comprises local energy generation amount, storage state and load demand, and the shared information from the adjacent micro-grids at least comprises marginal cost information representing energy supply and demand states of the adjacent micro-grids and global consistency variables of the adjacent micro-grids in a distributed consensus algorithm; s3, enabling a plurality of micro networks to be iteratively updated based on the shared information through a distributed consensus algorithm, and achieving agreement on a global optimization target, wherein the global optimization target is to minimize the total running cost of the cluster; And S4, updating the operation parameters of each micro-network according to the achieved consensus, and executing the optimized operation. Preferably, the iterative updating by the distributed consensus algorithm to agree on a global optimization target specifically includes: s31, constructing the global optimization target as a global optimization problem comprising a plurality of local variables of the micro-grid and at least one global coupling constraint, wherein the global coupling constraint is a switching power constraint of a cluster and