A power distribution network collaborative control and hardware-in-the-loop simulation system, method and device for new energy high proportion access configuration

Abstract

The application discloses a power distribution network collaborative control and hardware-in-the-loop simulation system, method and device for serving new energy high-proportion access configuration, relates to the technical field of power system testing and verification, and comprises a collaborative planning layer for constructing a source-load probability distribution model and generating a new energy configuration scheme based on a multi-objective optimization algorithm; an operation control layer for outputting steady-state indexes and transient response data under steady state and transient state and issuing control instructions according to the received new energy configuration scheme; a hardware-in-the-loop verification layer for constructing an electromagnetic transient closed-loop simulation environment, performing engineering-level dynamic verification on the control instructions issued by the operation control layer and feeding back real-time response data; and a dynamic display layer for performing multi-source fusion display and tracking of the new energy configuration scheme, the steady-state indexes and the transient response data through a unified time reference. The method realizes an integrated verification process of multi-source data acquisition, safe transmission, parameter issuing, control execution and result evaluation.

Description

Technical Field

[0001] This invention relates to the field of power system testing and verification technology, specifically to a distribution network collaborative control and hardware-in-the-loop simulation system, method, and apparatus for serving the high proportion of new energy access configuration. Background Technology

[0002] With the large-scale integration of distributed photovoltaic, wind power, and energy storage into the distribution network, the distribution network has gradually evolved from a traditional unidirectional power flow and weak power electronics characteristics into a complex system with multiple sources, multiple ports, and strong coupling. Furthermore, the power output of new energy sources is characterized by intermittency, volatility, and randomness, thus affecting the stable operation of the power grid. The mature development of hardware-in-the-loop (HIL) technology has been well-validated in multiple fields such as intelligent driving. This technology possesses high settlement capabilities and low-cost simulation and pre-simulation capabilities for complex operating conditions. Therefore, combining HIL technology with a distribution network collaborative control system configured with a high proportion of new energy integration is an inevitable trend. It is necessary to strengthen the research and development of key emergency technologies such as power grid disaster monitoring and early warning, and disaster response simulation, and to enhance the application of advanced technologies such as artificial intelligence to promote the construction of a smart emergency system and improve the level of technological support for power emergency response. This not only reaffirms the importance of simulation and artificial intelligence technologies for the development of smart grids but also further promotes the development of power grid system simulation verification. Currently, HIL testing often focuses on a single direction or a single scenario, lacking a comprehensive power system steady-state and power electronics transient collaborative control system. Therefore, a hardware-in-the-loop simulation test system is proposed to achieve multi-scenario, low-cost, switchable, and multi-parameter collaborative control of distribution networks with high proportion of new energy access.

[0003] Meanwhile, facing the challenge of accurately characterizing the dynamic characteristics of power devices and their equipment technologies under the trend of higher frequency and higher voltage in power systems, traditional testing methods struggle to cover the entire operating range, hindering the efficient application of novel power device topologies and control methods in power systems with a high proportion of power electronics. Furthermore, the planning and operation technology of high-proportion power electronic energy systems, as a crucial support for future smart grids, involves the deep integration of various types of power semiconductor devices with power systems. However, issues such as multi-timescale coupling, complex fault propagation mechanisms, and ambiguous operating boundaries significantly increase the difficulty of verifying system-level planning and operation control strategies. Moreover, for application testing in multiple scenarios, the construction of real-world testing environments not only greatly increases testing time and costs but also severely limits the testing and simulation of certain critical conditions. In summary, there is currently a lack of an integrated verification environment that can simultaneously achieve high precision, multi-physics coupling, and complex operating condition simulation. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is that existing methods for distribution networks in scenarios with a high proportion of new energy access have high costs for conducting panoramic experiments relying on different power devices, difficulties in verifying complex operating conditions, separation of steady-state operation assessment and power electronic transient control verification, and the problem of how to achieve unified modeling of test objects and control objects.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configurations, comprising: a collaborative planning layer, used to construct a source-load probability distribution model and generate renewable energy configuration schemes that balance steady-state economic indicators and transient safety indicators based on a multi-objective optimization algorithm; an operation control layer, which, based on the received renewable energy configuration schemes, performs energy management and outputs steady-state indicators in steady state, and models different renewable energy modules based on the characteristics of power electronic devices in transient state, and uses reinforcement learning algorithms to perform power electronic parameter self-tuning, output transient response data, and issue control commands; a hardware-in-the-loop verification layer, used to construct an electromagnetic transient closed-loop simulation environment, receive control commands issued by the operation control layer, perform engineering-level dynamic verification, and provide feedback real-time response data; and a dynamic display layer, which performs multi-source fusion display and tracking of renewable energy configuration schemes, steady-state indicators, and transient response data through a unified time base.

[0007] As a preferred embodiment of the distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion new energy access configuration described in this invention, the source-load probability distribution model is constructed by: considering the probability distribution model of wind power, photovoltaic power, and load uncertainty; using Weibull distribution to fit the wind speed frequency distribution characteristics, and using piecewise functions to describe the wind turbine power characteristic curve; using Beta distribution to describe the probability characteristics of solar irradiance, and combining ambient temperature to calculate the output power of the photovoltaic array; and using normal distribution to describe the load power demand characteristics in the distribution network.

[0008] As a preferred embodiment of the distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion access configuration of renewable energy described in this invention, the multi-objective optimization algorithm includes: using the configuration capacity of the components to be planned in the distribution network as decision variables, performing planning and solving in the solution space; sending the configuration scheme and scenario to the operation control module, which then calls the simulation calculation or hardware-in-the-loop verification and returns the calculation results; classifying the population through non-dominated sorting, maintaining population diversity by using congestion distance calculation, and outputting the Pareto optimal solution set as the renewable energy configuration scheme.

[0009] As a preferred embodiment of the distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration described in this invention, the operation control layer executes a hierarchical management strategy during steady-state operation to output steady-state indicators and power allocation to renewable energy sources; in transient response mode, a renewable energy topology module is constructed, transient safety indicators are output, and self-tuning of power electronic parameters is performed using a deep reinforcement learning algorithm. As a preferred embodiment of the distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration described in this invention, the steady-state energy management output includes, under steady-state conditions, performing energy management and power flow calculations based on the distribution network topology and source-load data, coordinating and optimizing the scheduling of distributed power sources, energy storage units, and load-side resources, allocating active and reactive power output, and outputting operating costs, power balance constraints, equipment capacity constraints, and voltage constraints, as well as outputting steady-state evaluation indicators for operating costs, renewable energy absorption rate, voltage qualification rate, and line current carrying rate.

[0010] As a preferred embodiment of the distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion access configuration of new energy described in this invention, the output transient response data includes: under transient conditions, constructing corresponding new energy topology modules according to the characteristics and application scenarios of different power devices, completing the coupling connection between modules of different energy sources, generating control commands according to the disturbance scenario and driving the lower-level real-time simulation environment to obtain dynamic response data of frequency and voltage.

[0011] As a preferred embodiment of the distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration described in this invention, the following steps are included: receiving control commands issued by the operation control layer, performing engineering-level dynamic verification, and feeding back real-time response data. This includes: constructing a distribution network electromagnetic transient closed-loop simulation environment based on steady-state and transient control analysis, at a real time scale and combined with the characteristics of real power devices, to perform engineering-level dynamic verification of the collaborative control strategy; using the RT-LAB real-time simulation platform as the core operating carrier, building an electromagnetic transient model in the real-time simulator that includes distributed renewable energy grid-connected interfaces, energy storage units, and load models; performing numerical calculations according to a fixed simulation step size, and outputting node voltage, current, frequency, and related state variables in real time; and connecting control commands generated by the operation control layer to the RT-LAB system through a physical interface, with RT-LAB updating the model state and feeding back dynamic response data based on the control signals.

[0012] As a preferred embodiment of the distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion access configuration of new energy described in this invention, the dynamic display layer performs visual monitoring and comparative analysis of collaborative planning results and key parameters of power electronic control; and constructs a data acquisition, industrial communication and secure transmission, data upload and platform display process for the distribution network collaborative control and hardware-in-the-loop simulation system.

[0013] Another objective of this invention is to provide a distribution network collaborative control and hardware-in-the-loop simulation method for high-proportion renewable energy access configurations. This method can construct an electromagnetic transient closed-loop simulation environment, perform engineering-level dynamic verification of control commands issued by the operation control layer, and provide a hardware-in-the-loop verification layer that feeds back real-time response data. This solves the problem of difficulty in unified modeling in current distribution network methods for high-proportion renewable energy access configuration scenarios.

[0014] As a preferred embodiment of the distribution network collaborative control and hardware-in-the-loop simulation method for high-proportion renewable energy access configuration described in this invention, the method includes: constructing a source-load probability distribution model based on collected basic data of the distribution network, generating a steady-state scenario set and a transient scenario set; constructing a multi-objective optimization configuration model that takes into account both steady-state and transient indicators, generating an initial renewable energy configuration scheme, and distributing the initial renewable energy configuration scheme and scenario set to the operation control layer; constructing a corresponding renewable energy control module in the operation control layer, and performing optimization iteration of control parameters based on a reinforcement learning model to ensure good follow-up of the issued command parameters; interacting with the simulation environment using corresponding coupling interfaces to generate a chained real-time control strategy; distributing the real-time control strategy and corresponding control commands to the hardware-in-the-loop verification layer to drive the simulator to run the distribution network electromagnetic transient model, and collecting response data through physical interfaces using hardware collaborative devices; outputting transient safety indicators based on the feedback dynamic response data, and combining the energy management performed by the operation control layer to obtain steady-state indicators and iterating until the optimal configuration scheme set is output.

[0015] Another object of the present invention is to provide a distribution network collaborative control and hardware-in-the-loop simulation device for serving a high proportion of new energy access configuration, wherein a computer program is stored thereon, and when the computer program is executed by a processor, the steps of the distribution network collaborative control and hardware-in-the-loop simulation system for serving a high proportion of new energy access configuration are implemented.

[0016] The beneficial effects of this invention are: This invention provides a distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configurations. This system constructs a source-load probability distribution model and generates a collaborative planning layer based on a multi-objective optimization algorithm to achieve renewable energy configuration schemes that balance steady-state economic indicators and transient safety indicators. It addresses uncertainty through probabilistic modeling, balances economy and safety through multi-objective optimization, and ensures the feasibility of results through interaction with lower-level hardware, providing precise configuration schemes for distribution networks with high renewable energy access ratios. Based on the received renewable energy configuration scheme, the system performs energy management and outputs steady-state indicators in steady state. In transient state, it utilizes reinforcement learning algorithms to perform power electronic parameter self-tuning, outputs transient response data, and issues control commands to an operation control layer. In transient state, a corresponding renewable energy control module is constructed in the operation control layer to effectively follow the issued command parameters. Furthermore, it innovatively introduces reinforcement learning algorithms to achieve adaptive online tuning of the PI parameters of the power electronic controller, solving the problem of adaptability issues in traditional fixed-parameter control in high-proportion renewable energy scenarios. This invention addresses the challenges of poor responsiveness and low efficiency of manual parameter tuning. It constructs an electromagnetic transient closed-loop simulation environment, performing engineering-level dynamic verification of control commands issued by the operation control layer and feeding back real-time response data. This hardware-in-the-loop verification layer interacts with the real controller through a physical interface, enabling low-cost and repeatable simulation of various complex faults and critical conditions without the need for a fully physical experimental platform. It accurately captures dynamic indicators such as frequency extrema and voltage response. A dynamic display layer, using a unified time base, integrates and tracks multi-source data from new energy configuration schemes, steady-state indicators, and transient response data. This layer aligns and integrates upper-layer planning schemes, mid-layer steady-state indicators, and lower-layer hardware-in-the-loop dynamic data such as voltage and frequency, achieving multi-dimensional comparative analysis and curve overlay display from configuration schemes to control effects. This invention achieves superior results in integrated collaborative control, intelligent adaptive parameter control, high-fidelity hardware-in-the-loop dynamic verification, and full-process traceable evaluation. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a module functional logic and data flow structure diagram of a distribution network collaborative control and hardware-in-the-loop simulation system for serving high proportion of new energy access configuration provided in Embodiment 1 of the present invention.

[0019] Figure 2 This is a flowchart of the collaborative planning layer of a distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion access configuration of renewable energy provided in Embodiment 1 of the present invention.

[0020] Figure 3 This is a schematic diagram of the PI parameter self-tuning process based on reinforcement learning in the HIL of a distribution network collaborative control and hardware-in-the-loop simulation system for a high proportion of new energy access configuration provided in Embodiment 1 of the present invention.

[0021] Figure 4 This is a schematic diagram of the hardware-in-the-loop collaborative control hardware device architecture of a distribution network collaborative control and hardware-in-the-loop simulation system for serving high proportion of new energy access configuration provided in Embodiment 1 of the present invention.

[0022] Figure 5 This is a schematic diagram illustrating the process of uploading hardware-in-the-loop test data to the platform for a distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion access configuration of new energy sources, as provided in Embodiment 1 of the present invention. Detailed Implementation

[0023] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0024] Example 1, referring to Figures 1-5 As an embodiment of the present invention, a distribution network collaborative control and hardware-in-the-loop simulation system for serving high-proportion renewable energy access configurations is provided, comprising: S1: Collaborative planning layer 100, used to construct source-load probability distribution model 101, and generate new energy configuration scheme 102 that takes into account both steady-state economic index T1 and transient security index T2 based on multi-objective optimization algorithm.

[0025] Specifically, it is used to collect basic data of the distribution network, construct a source-load probability distribution model 101, generate steady-state typical scenarios and transient disturbance scenarios, establish a configuration optimization model that takes into account both steady-state performance and transient performance, and generate a new energy configuration scheme 102 based on a multi-objective optimization algorithm. This scheme is then sent to the operation control layer 200 as configuration logic to realize the planning and solution of the new energy access capacity of the distribution network.

[0026] The source-load probability distribution model 101 includes a probability distribution model that considers the uncertainties of wind power, photovoltaics, and load; a Weibull distribution is used to fit the wind speed frequency distribution characteristics, and a piecewise function is used to describe the wind turbine power characteristic curve; a Beta distribution is used to describe the probability characteristics of solar irradiance, and the output power of the photovoltaic array is calculated in combination with the ambient temperature; and a normal distribution is used to describe the load power demand characteristics in the distribution network.

[0027] The multi-objective optimization algorithm includes: using the configuration capacity of the components to be planned in the distribution network as the decision variable, performing planning and solving in the solution space; sending the configuration scheme and scenario to the operation control module, which then calls simulation calculation or hardware-in-the-loop verification and returns the calculation results; classifying the population through non-dominated sorting, maintaining population diversity by using crowding distance calculation, and outputting the Pareto optimal solution set as the new energy configuration scheme 102.

[0028] Furthermore, the steady-state typical scenarios are a set of long-period typical scenarios with a time scale of 15 minutes or more. Latin hypercube sampling (LHS) is used to perform stratified sampling in the source-load variable space, and the K-means clustering algorithm is used to reduce the samples into several steady-state typical scenarios, ultimately forming a set of steady-state scenarios.

[0029] Transient disturbance scenarios are sets of transient disturbance scenarios with a small time scale that include information about the beginning and end of state transitions. Using Gaussian Copula theory, a conditional normal distribution between the current state and the state at the next moment is established using the spatiotemporal correlation coefficient matrix. This is used to sample and generate a set of transient disturbance scenario vector pairs that include load jumps and fluctuations in renewable energy power.

[0030] The configuration optimization model that balances steady-state and transient performance comprises a multi-objective system consisting of steady-state and transient indices. As a preferred embodiment of this invention, the steady-state indices include distribution network operating costs and renewable energy absorption rate, while the transient indices include maximum frequency deviation, and also include several transient constraint measures.

[0031] It should be noted that, firstly, the original topology of the distribution network, historical meteorological data, and load characteristics are collected to construct a source-load probability distribution model 101. Secondly, typical steady-state scenarios are generated using Latin hypercube sampling (LHS), and transient disturbance scenarios including load jumps and new energy power fluctuations are generated using Gaussian Copula theory. Subsequently, a multi-objective optimization model that considers both steady-state and transient indicators is constructed, and the NSGA-II algorithm is used to search for the configuration capacity of distributed power sources in the solution space. This layer does not directly calculate specific operating indicators; instead, it distributes its corresponding operating scenarios to the operation control layer 200 and the hardware-in-the-loop verification layer 300. The generated configuration scheme and corresponding scenarios are then "distributed" to the operation control layer 200, where the fitness value is fed back from the lower layer to guide the continuous iteration of the configuration scheme.

[0032] Accurately describing the stochastic characteristics of wind power, solar power, and loads is fundamental for subsequent scenario generation and optimization. The following describes the stochastic characteristics of wind power and solar power loads at collaborative planning layer 100. The randomness of wind speed is the fundamental source of fluctuations in wind power generation. A Weibull distribution is used to fit the frequency distribution characteristics of actual wind speed, and its probability density function is expressed as: , in, Let be the probability density function. The shape parameter of the Weibull distribution reflects the width of the wind speed distribution. The scale parameter of the Weibull distribution reflects the average wind speed level. Wind speed (m / s).

[0033] Because there is a non-linear mapping relationship between the output power of a wind turbine and wind speed, the power characteristic curve of the wind turbine can be represented by a piecewise function as follows: , in, This refers to the output power of the wind turbine. To cut into wind speed, To cut off the wind speed, This refers to the rated power of the fan. This is the rated wind speed.

[0034] The output of photovoltaic power generation mainly depends on solar irradiance. A Beta distribution is used to describe the probabilistic characteristics of irradiance over a specific time period, simulating the effects of factors such as cloud cover and atmospheric transmittance. Irradiance The probability density function is expressed as: , in, Irradiance The probability density function, For the Gamma function, Let be the shape parameter of the Beta distribution. Irradiance, This represents the maximum possible irradiance.

[0035] The output power of a photovoltaic array is approximately linearly related to irradiance and is affected by ambient temperature. The specific relationship between the output power of a photovoltaic array and irradiance and temperature is expressed as follows: , in, The output power of the photovoltaic array. For the area of ​​the photovoltaic panel, For photoelectric conversion efficiency, This refers to the power temperature coefficient of the battery. Battery operating temperature This is a reference temperature (usually 25°C).

[0036] Load demand in a distribution network is influenced by user behavior, weather conditions, and economic activities. A normal distribution is used to describe the load characteristics of a distribution network, meaning the load power in the distribution network satisfies the following expression: , in, The load power in the distribution network, This is the average load forecast. The standard deviation is denoted as .

[0037] It should also be noted that, in order to evaluate the steady-state indicators of the distribution network (such as operating costs and absorption rate) during the planning stage, it is necessary to generate representative steady-state long-term scenarios. This is achieved using a method combining Latin hypercube sampling (LHS) and K-means clustering. LHS is a stratified sampling technique whose core idea is to divide the probability distribution interval of each input variable (wind, solar, load) into... The method divides the data into equal-probability sub-intervals and randomly selects a sample from each sub-interval. The physical value is then obtained by mapping the data using the inverse function of the cumulative distribution function. This method ensures that the sample points uniformly cover the entire sample space.

[0038] Assuming it is formed by wind, light, and load. random variables It needs to generate The first sample. Variables , and its cumulative distribution function (CDF) range Divided into There are three non-overlapping intervals of equal width, each with a width of [value missing]. .make In the each interval Generate a random number that follows a uniform distribution. As shown in the equation, the inverse function of CDF Calculate the corresponding sample values.

[0039] , in, For sample values, It is the inverse function of CDF. These are random numbers that follow a uniform distribution.

[0040] LHS generated sample size The data is still too large. To improve computational efficiency, clustering algorithms are needed to extract typical scenes. The K-means algorithm is chosen for scene clustering to form scene sets. This algorithm is widely used for scene reduction due to its simplicity and efficiency. K-means aims to... The scene samples are divided into: Find clusters such that the sum of squared errors within each cluster is minimized, i.e., minimize the objective function. , is represented as: , in, To minimize the objective function, For the first The sample set of each cluster For scene vectors, For the first The centroid of a cluster.

[0041] After clustering, the centroid of each cluster is... This is a typical scenario, and the probability of this scenario occurring. The number of original samples contained within the cluster A decision is expressed as: , in, This represents the probability of the scenario occurring. The original sample size. This represents the total number of scene samples.

[0042] The final set of steady-state scenarios is represented as follows: , in, This is a set of typical steady-state scenarios. For the first The centroid of a cluster, This represents the probability of the scenario occurring.

[0043] It should also be noted that with a high proportion of renewable energy integration, instantaneous power jumps between source and load may trigger frequency exceedances. To assess the transient safety of the planning scheme, a set of transient scenarios containing state transition information must be generated, and this set is constructed based on Gaussian Copula theory. The construction of transient scenarios is essentially a conditional simulation problem of a high-dimensional stochastic process, i.e., given the system state at a certain moment... Under these conditions, by taking into account the spatiotemporal correlation between variables, sampling is generated. System state at time 1 This forms a vector pair containing state transition information. The dataset is then condensed. Subsequently, the massive number of scenarios is reduced to meet the efficiency requirements of subsequent transient simulation calculations.

[0044] definition System state vector at time t This vector consists of the wind power output, photovoltaic power output, and load demand of the distribution network, where... The total dimension of the variable.

[0045] Calculate the linear correlation coefficient matrix Based on historical operation data of the power distribution network and meteorological time series data, node variables at adjacent time points are extracted and spliced ​​to construct a joint data matrix. To eliminate the distortion of correlation assessment caused by nonlinear probability integral transformation, the Kendall rank correlation coefficients of the joint matrix elements are calculated. To construct the rank correlation matrix Subsequently, the parsing mapping relationship was utilized. The linear correlation coefficient matrix required to convert the rank correlation matrix into a Gaussian Copula. .

[0046] Using variables The cumulative probability distribution function and the inverse standard normal distribution function map the current state vector in the physical space to a vector in the standard normal space, expressed as: , in, This is a vector that maps the current state vector to the standard normal space. The cumulative distribution function of the standard normal distribution. For the first Marginal cumulative distribution function of variables, For the first The variables at time... The original physical values.

[0047] In normal space, joint vector Follows a multivariate standard normal distribution ,in Spatiotemporal correlation coefficient matrix. The correlation coefficient matrix... Divided by time block , and .

[0048] Based on the conditional probability property of the multivariate normal distribution, given the current state... The state at the next moment Follows a conditional normal distribution Its conditional mean Conditional covariance The calculation method is expressed as follows: , , in, For conditional mean, For conditional covariance, The cross-correlation matrix, It is the inverse of the autocorrelation matrix. Let this be the current state vector. For the future time autocorrelation matrix, This is the transpose of the cross-correlation matrix.

[0049] Introducing independent standard normal random variables And the conditional covariance matrix was decomposed using Cholesky decomposition to obtain Then the normal sample at the next time step The generation is represented as: , in, For the next time step, a normal sample. Decompose the lower triangular matrix using the Cholesky method. It is a normally distributed random variable.

[0050] Finally, through inverse transformation By restoring the sample to physical space, a transient scene sample can be obtained. Using the Monte Carlo method, the above process is repeated cyclically to generate... A transient scenario sample A set of transient scenario samples is formed.

[0051] A scene reduction technique is used to extract the most representative transient scenes. First, two transient scenes are defined. and The distance cost between them is expressed as: , in, For transient scenarios and Distance cost between For the scene exist The state vector at time t, For the scene exist The state vector at time t, For the scene exist The state vector at time t, For the scene exist The state vector at time t.

[0052] The Fast Forward Selection (FFS) algorithm is used to iteratively filter typical scenarios. The core idea of ​​this algorithm is to find a subset. To make it consistent with the original scene set probability distance between Minimize. The final output contains The set of typical vector pairs and their corresponding probability weights constitutes the transient scenario set. .

[0053] In practical planning, the allocation capacity planning of renewable energy in distribution networks should not only consider steady-state indicators such as economic efficiency, but also the transient security of the distribution network under scenarios with a high proportion of renewable energy access. In this invention, the planning model is no longer limited to a single economic indicator, but rather a comprehensive planning model covering both steady-state economics and transient security. Minimizing operating costs, minimizing renewable energy curtailment rate, and minimizing maximum frequency deviation are selected as optimization objectives, with frequency transient constraints as constraints, to construct a multi-objective optimization model.

[0054] Minimize annualized operating cost: Minimize the total annualized cost over the entire lifecycle, including annualized investment cost. Annualized operation and maintenance costs in typical scenarios Represented as: , in, To minimize annualized operating costs, This refers to the annualized investment cost of distributed power sources calculated based on the planned capacity. This is a set of typical steady-state scenarios. For the scene The probability of occurrence, This refers to the daily operation and maintenance cost fed back from the operation control layer 200 in this scenario.

[0055] Minimize the curtailment rate of renewable energy: Maximize the absorption rate of renewable energy, i.e., minimize the curtailment rate of wind and solar power. It is defined as the ratio of total power generation to the amount of curtailed power, expressed as: , in, It is the ratio of the difference between total power generation and the amount of power wasted. This is a set of typical steady-state scenarios. For the scene Next moment The theoretical maximum power generation capacity of new energy sources This represents the actual power received by the operation control layer 200 after optimized scheduling.

[0056] Minimizing the maximum frequency deviation is expressed as: , in, To minimize the maximum frequency deviation, This is a set of transient disturbance scenarios. In disturbed scenarios Next, the system frequency is fed back after the hardware-in-the-loop simulation is called by the operation control layer 200 and the hardware-in-the-loop verification layer 300. The rated frequency is 50Hz.

[0057] For all transient scenarios, transient safety constraints are specifically introduced, including the minimum (or maximum) frequency constraint and the RoCoF constraint on the rate of change of frequency.

[0058] The constraint at the lowest (or highest) frequency point is expressed as: , in, The lower limit requirement for frequency in the distribution network. This refers to the upper limit requirement for frequency in the power distribution network.

[0059] The RoCoF constraint on the rate of change of frequency is expressed as: , in, For the perturbation scene The rate of change of frequency, The maximum acceptable rate of frequency change.

[0060] The above planning model is a typical nonlinear, nonconvex, multi-objective optimization problem (MINLP). Traditional weighted summation methods are difficult to handle the nonconvex Pareto front between objectives.

[0061] This invention employs a non-dominated sorting genetic algorithm (NSGA-II) with an elitist strategy to solve the problem. Through interactive iteration with the running control layer 200, it outputs a Pareto optimal solution set. The specific steps are as follows: Define the optimization variable vector. ,in and Representing the first The first wind turbine node and the first The configuration capacity of each photovoltaic node. The population size is set to... The maximum number of iterations is Within the feasible region of the decision variables Initial population is generated randomly within the population. Each individual in the population This represents a potential new energy configuration scheme 102. Simultaneously, the generated steady-state scenario set is loaded. With transient scene set For each individual in the population (i.e., the first) (Several configuration options) Execute the following evaluation process: This configuration scheme is combined with the steady-state scenario set. The data is then sent to the Operation Control Layer 200. The Operation Control Layer 200 performs steady-state optimal power flow calculations, feeds back the actual energy management scheme, and provides it to the Collaborative Planning Layer 100 for calculating the objective function. and .

[0062] This configuration scheme and transient scenario set The command is sent to the operation control layer 200. The operation control layer 200 generates control commands and drives the hardware to perform dynamic simulation in the loop layer, capturing the dynamic frequency trajectory and feeding back the frequency extreme values. With rate of change Provided for the collaborative planning layer 100 to calculate the objective function And check the constraints.

[0063] Based on the calculated three objective function values This involves ranking the population using non-dominated sorting. Individuals are defined. Dominant Individual (recorded as) ), holds if and only if the expression is satisfied.

[0064] , in, For individuals The The objective function value, For individuals The The objective function value, For individuals The The objective function value, For individuals The The objective function value.

[0065] Based on dominance relationships, the solution set that is not dominated by any other individual in the population is selected to form the Pareto front hierarchy. Remove from the population Then, the non-dominated solution set of the remaining individuals is obtained. Similarly, based on dominance relationships, the population is divided into several Pareto front levels. ,in This is the optimal layer (non-dominated solution set). Individuals in the layer are only to Individual dominance within the layer.

[0066] To maintain the diversity of solution distribution in the target space, for the same level Individual crowding distance calculation For the first Each objective function is used to sort the individuals in this layer in ascending order of their objective values. The congestion distance is defined as follows: , in, For individuals Crowding distance The first adjacent individual after sorting The objective function value, This represents the maximum value of the objective function at the current level. This is the minimum value of the objective function at the current level.

[0067] A binary tournament selection process based on rank and crowding is used to retain high-quality individuals. A simulated binary crossover (SBX) and polynomial mutation operator are used to generate the offspring population. Parents and offspring are merged, and non-dominated sorting and crowding selection are re-executed to produce a new generation. The above steps are repeated until the preset maximum number of iterations is reached. After the algorithm terminates, it outputs the Pareto optimal frontier. Planners can then select the final distribution network renewable energy configuration scheme 102 from the optimal solution set based on actual engineering preferences.

[0068] S2: Operation control layer 200, based on the received new energy configuration scheme 102, performs energy management and outputs steady-state indicators in steady state, and uses reinforcement learning algorithm to perform power electronic parameter self-tuning in transient state, outputs transient response data and issues control commands.

[0069] Specifically, the operation control layer 200 is the core of the system's operation calculation and control decision-making. It undertakes three functions: steady-state index calculation, transient control strategy generation, and optimization feedback. This layer receives the new energy configuration scheme 102 and operation scenarios sent down from the upper layer.

[0070] The steady-state energy management output steady-state indicators include: under steady-state conditions, energy management and power flow calculations are performed based on the distribution network topology and source-load data; coordinated and optimized scheduling of distributed power sources, energy storage units and load-side resources are carried out; active and reactive power outputs are allocated; and conditions such as operating costs, power balance constraints, equipment capacity constraints and voltage constraints are taken into account in a comprehensive manner to achieve priority consumption of new energy and economic operation of the system. The steady-state evaluation indicators of operating costs, new energy consumption rate, voltage qualification rate and line current carrying rate are output, providing a quantitative basis for multi-objective optimization at the planning level.

[0071] The output transient response data includes, under transient conditions, constructing corresponding new energy topology modules according to the characteristics and application scenarios of different power devices, completing the coupling connection between modules of different energy sources, generating control commands according to the disturbance scenario and driving the lower-level real-time simulation environment, obtaining dynamic response data of frequency and voltage, calculating transient safety indicators and feeding them back to the planning layer for optimization iteration.

[0072] Furthermore, the operation control layer 200 incorporates a reinforcement learning-based PI parameter self-tuning mechanism. This mechanism constructs a state, action, and reward mapping model to achieve adaptive optimization of the PI parameters. The state is characterized by real-time observations under the current simulation topology, as well as deviations between reference and feedback values. Actions are defined as dynamic adjustments to the proportional and integral coefficients of the power electronic controller based on the current operating state using reinforcement learning. Specifically, this involves dynamically increasing, decreasing, or maintaining the parameters according to preset rules. The reward function employs a multi-objective comprehensive evaluation approach, applying positive incentives or negative penalties based on indicators such as the system's deviation from the set target and transient oscillation amplitude, guiding the algorithm to search for and determine the optimal PI parameters during iteration. This mechanism effectively ensures rapid response and stable operation of the system during transient processes, enabling adaptive optimization and closed-loop verification of power electronic control parameters. This, in turn, improves the stability and engineering feasibility of the control strategy under conditions of high-proportion renewable energy access.

[0073] Furthermore, the operation control layer 200 serves as the system's computational core. This layer receives the new energy configuration scheme 102 and operating scenarios from the top layer. Under steady-state logic, this layer executes energy management algorithms to calculate indicators such as distribution network operating costs and new energy absorption rates. Under transient logic, it defines the system's control strategy when facing disturbances and forwards control commands to the underlying hardware-in-the-loop verification layer 300. In the hardware-in-the-loop simulation environment, it ultimately calculates transient safety indicators based on the voltage and frequency dynamic curves fed back from the bottom layer and sends all calculation results back to the collaborative planning layer 100 for evaluation.

[0074] First, the core of energy management algorithms is optimization algorithms, which are typically implemented by writing code based on a planned scheme to manage energy.

[0075] After receiving the new energy configuration scheme 102 and corresponding operating scenarios from the collaborative planning layer 100, the operation control layer 200 first parses the configuration parameters and scenario data and maps them into the distribution network operation model, completing the initialization of network topology, source-load parameters, and equipment constraints. Secondly, based on typical steady-state scenarios, it constructs an energy management optimization model, comprehensively considering power flow balance, equipment capacity, voltage constraints, and energy storage operating boundaries, to solve for the optimal scheduling sequence of distributed power sources, energy storage, and flexible loads, thus forming an energy management scheme for the corresponding scenario. Subsequently, based on this scheme, it calculates steady-state indicators such as operating costs and new energy absorption rate, and further combines transient disturbance scenarios to generate dynamic control strategies, which are then sent to the hardware-in-the-loop verification layer 300 to obtain dynamic response results such as voltage and frequency, and then calculate transient safety indicators. Finally, the operation control layer 200 feeds back the steady-state and transient evaluation results to the collaborative planning layer 100 as the basis for judging the merits of the configuration scheme and subsequent iterative optimization.

[0076] It should be noted that the new energy configuration scheme 102 and operation scenario issued by the collaborative planning in the collaborative planning layer 100, as the operation control layer 200, consists of two parts: an optimized scheduling model and a real-time control strategy. It is the core of operation decision-making connecting the planning layer and the hardware-in-the-loop verification layer 300. In the optimized scheduling model part, the operation control layer 200 constructs a distribution network optimized scheduling model under steady-state logic. It comprehensively considers source-load balance constraints, equipment capacity constraints, voltage constraints, and operation economic objectives, and coordinates and optimizes the scheduling of distributed power sources and energy storage resources. It performs energy management calculations to achieve reasonable allocation of active and reactive power and priority consumption of new energy, and obtains steady-state indicators such as operating costs, consumption rate, and power flow distribution, providing quantitative basis for model solving and scheme selection in the planning layer. In the real-time control part of new energy power electronics, the operation control layer 200 first constructs corresponding new energy power electronics topology modeling and coupling interfaces of each module based on the operating characteristics and application scenarios of different power devices, so as to support the realization of dynamic and stable operation of new energy modules in transient states and good tracking of stable control. Subsequently, the control strategy model is trained and generated using methods such as reinforcement learning. Combined with the distribution network's operating status and disturbance scenarios, a real-time control parameter tuning iterative scheme is formed, optimizing traditional power electronic parameter tuning and laying the foundation for more stable and faster dynamic control tracking. Under transient logic, this layer outputs distribution network transient state settings and control commands, which are sent to the hardware-in-the-loop verification layer 300 for execution. It also receives real-time response data such as distribution network voltage and frequency, evaluates transient operating waveforms and dynamic performance indicators, and achieves closed-loop linkage between operation control and dynamic verification. Therefore, the operation control layer 200, through the coordinated operation of steady-state optimized scheduling and transient real-time control, supports the multi-objective optimization iteration of the planning layer on the one hand, and ensures the feasibility and stability of the control strategy under actual dynamic operating conditions on the other. It is the core decision-making and coordination unit at the system operation level.

[0077] Simultaneously, in transient control, a corresponding new energy control module is constructed at the operational control layer, and reinforcement learning is used to optimize and iteratively adjust the PI parameters in the control strategy. For example... Figure 3 As shown, in the transient control of the operation control layer 200, to meet the collaborative control requirements of the distribution network serving the high proportion of new energy access, a combination of simulation environment and reinforcement learning is used to complete the PI parameter self-tuning of the power electronic simulation control module of a single new energy module. This provides a stable, reproducible, and constraint-satisfied set of control parameters for subsequent collaborative control closed-loop operation. The operation control layer 200 steps are as follows: Figure 3 The process shown executes the following steps sequentially.

[0078] First, the startup initialization steps are executed to initialize the power electronics simulation control module, construct the corresponding new energy control module and interface coupling, and bring it into a tuneable state. This includes clearing basic variables, checking control loop enable conditions, and preparing communication and sampling channels with the simulation system. Next, the basic configuration steps of the simulation system are executed to configure the topology, operating conditions, disturbance methods, and observations of the simulation object, and to limit the feasible range and constraints of the PI parameters. These constraints include at least one or more of the following: parameter upper and lower bounds, control quantity limits, response speed constraints, and overshoot constraints, to ensure that subsequent learning and output parameters meet engineering feasibility and safety requirements.

[0079] After completing the basic configuration of the simulation system, the reinforcement learning model building step is executed to construct a reinforcement learning model for PI parameter mapping. This model enables the reinforcement learning model to output PI control parameters or their increments based on system state information. The input states of the reinforcement learning model include at least one or more of the following: voltage or current deviation, deviation rate of change, power fluctuation, control output, and system operating mode identifier. The action space of the reinforcement learning model includes at least the values ​​or adjustments of the proportional and integral coefficients. Further, the reinforcement learning model training and optimization steps are executed, allowing the model to adaptively optimize the PI parameters through interaction with the simulation environment.

[0080] During training and optimization, the boundary and reward / penalty design steps are first executed to define the environment, state, and action space, and set the PI parameter range and constraint relationships. Simultaneously, a reward / penalty function is constructed to guide the learning direction. This function considers at least the steady-state error, dynamic overshoot, steady-state fluctuation error, and the amplitude of control variable changes, and weights can be assigned to these indicators according to different operating conditions to adapt to the multi-objective control requirements under conditions of high-proportion renewable energy access. Subsequently, iterative optimization and stopping criterion steps are executed. Through repeated interactive sampling and updating of the policy network or value network with the simulation environment, the cumulative reward gradually increases and tends to stabilize. When the cumulative reward reaches a stable threshold, or the error index reaches a preset standard, or the number of iterations reaches the upper limit, the current round of training stops, and the current optimal parameters are output.

[0081] Once the stopping criterion is met, the output PI control parameter step is executed. The PI control parameters obtained from the reinforcement learning model are output as candidate tuning results and written into the parameter set for simulation closed-loop verification. Next, the simulation system operation and feedback acquisition step is executed. The PI parameters are loaded into the control module, and the simulation system is run under set operating conditions and disturbance conditions. Relevant parameters of the control module are collected as feedback information. The feedback information includes at least one or more of the following: output response curve, steady-state error statistics, overshoot, settling time, and number of control output limiting triggers. Subsequently, the monitoring and optimization iteration step is executed. The feedback parameters are evaluated and recorded, and the evaluation results are used to determine whether further parameter optimization is needed.

[0082] In the parameter acceptance phase, a control output response satisfaction judgment step is executed, determining whether the control output response meets the requirements based on preset evaluation indicators. These evaluation indicators include at least one or more of the following conditions: steady-state error does not exceed a threshold, overshoot does not exceed a threshold, settling time does not exceed a threshold, control output does not experience continuous saturation, and stability is maintained under typical disturbances. If the judgment result is unsatisfactory, the process returns to the training and optimization steps, adjusting reward / penalty weights, constraint boundaries, or training strategies based on feedback information, and continuing iterative optimization to achieve closed-loop self-tuning of the PI parameters. Finally, if the judgment result is satisfactory, the process ends, the tuned PI parameters are put into operation and permanently saved, putting the system into closed-loop operation, providing a directly callable control parameter foundation for subsequent collaborative control and hardware-in-the-loop verification phases.

[0083] S3: Used to build an electromagnetic transient closed-loop simulation environment, and to perform engineering-level dynamic verification of control commands issued by the operation control layer 200 and to provide feedback on real-time response data in the hardware-in-the-loop verification layer 300.

[0084] Specifically, the hardware-in-the-loop verification layer 300 includes external hardware devices or apparatuses such as real-time simulators and physical controllers. Its core function is to construct a closed-loop electromagnetic transient simulation environment for power distribution networks under real time scales and in combination with the characteristics of real power devices. This includes the characteristics of power devices and the multi-scale behavior of semiconductor power systems, and to perform engineering-level dynamic verification of collaborative control strategies.

[0085] The process of performing engineering-level dynamic verification of control commands issued by the operation control layer 200 and feeding back real-time response data includes: based on steady-state and transient control analysis, constructing a closed-loop electromagnetic transient simulation environment for the distribution network under real time scale and combined with the characteristics of real power devices, and performing engineering-level dynamic verification of the collaborative control strategy; using the RT-LAB real-time simulation platform as the core operating carrier, building an electromagnetic transient model in the real-time simulator that includes distributed new energy grid-connected interfaces, energy storage units, and load models; performing numerical calculations according to a fixed simulation step size, and outputting node voltage, current, frequency, and related state variables in real time; and connecting the control commands generated by the operation control layer 200 to the RT-LAB system through a physical interface, with RT-LAB updating the model state and feeding back dynamic response data based on the control signals.

[0086] Furthermore, the real-time simulator is used to run the electromagnetic transient model of the distribution network, which includes distributed new energy grid connection interfaces, energy storage units and load models, and outputs the voltage, current and internal state quantities of each node and the system according to the preset simulation step size and numerical compensation mechanism, so as to ensure the numerical stability and dynamic response accuracy under real-time calculation conditions.

[0087] The physical controller interacts with the real-time simulator and external circuits or terminal equipment through analog interfaces, digital interfaces or industrial communication interfaces, outputs control signals to the electromagnetic transient model, and receives model feedback signals to form a complete closed-loop control circuit.

[0088] During closed-loop operation, the hardware-in-the-loop verification layer 300 not only achieves high-speed operation of electromagnetic transient simulation, but also continuously collects and transmits key response data of the distribution network in real time. The data includes at least the voltage amplitude and phase angle of key nodes, system frequency and frequency change rate, output power of power electronic devices and control state quantities, etc. Time alignment is performed through a unified time base, providing real time-scale data support for the dynamic performance evaluation of operation control strategies and the calculation of transient safety indicators, thereby realizing the verification and evaluation of collaborative control schemes under engineering application conditions.

[0089] It should be noted that the Hardware-in-the-Loop (HIL) verification layer 300 is located at the bottom of the architecture. This layer consists of an RT-LAB real-time simulator and external hardware such as physical controllers. The HIL verification layer 300 constructs a closed-loop electromagnetic transient simulation environment for the power distribution network at a real-time scale, combining the characteristics of real power devices, to perform engineering-level dynamic verification of the collaborative control strategy. The RT-LAB real-time simulator runs a high-precision electromagnetic transient model of the power distribution network and connects to the power electronic controller via physical boards. The controller senses the model state through the physical boards and executes the strategies issued by the control layer. The RT-LAB real-time simulator runs the high-precision electromagnetic transient model of the power distribution network and transmits the distribution network voltage and frequency signals back to the model in real time.

[0090] It should also be noted that the core function of the hardware-in-the-loop verification layer 300 is based on the steady-state and transient control analysis of the operation control layer 200. Under real-time scales and combined with the characteristics of real power devices, it constructs a closed-loop electromagnetic transient simulation environment for the distribution network, enabling engineering-level dynamic verification of collaborative control strategies. This layer uses the RT-LAB real-time simulation platform as its core operating carrier. An electromagnetic transient model, including distributed new energy grid-connected interfaces, energy storage units, and load models, is built in the real-time simulator. Numerical calculations are performed according to a fixed simulation step size, and node voltages, currents, frequencies, and related state variables are output in real time. The control commands generated by the operation control layer 200 are connected to the RT-LAB system through a physical interface. RT-LAB updates the model state based on the control signals and feeds back dynamic response data, thus forming a real-time closed-loop verification environment between the control strategy and the electromagnetic transient model. Through this layer, key indicators such as frequency extrema, rate of change, voltage response, and power dynamics can be obtained, providing high-precision data support for transient safety assessment and control performance analysis.

[0091] Building upon this foundation, a supporting collaborative control hardware device was designed to ensure the stable operation and engineering adaptability of the RT-LAB hardware-in-the-loop verification process. This device primarily handles signal acquisition, control output, communication interaction, and data recording with RT-LAB, including modules for data acquisition, clock synchronization, control calculation, interface communication, and log storage. It ensures real-time execution of control commands and reliable transmission of feedback data. This support hardware is not an independent functional layer but rather serves as an interface and operational support unit for the RT-LAB real-time simulation environment, enhancing the real-time performance, stability, and traceability of hardware-in-the-loop verification.

[0092] like Figure 4 As shown, in the hardware-in-the-loop verification layer 300, based on the scenario data output from the preceding steps, the description of the collaborative control task, the point-table mapping relationship, and the interface configuration parameters, a hardware-in-the-loop collaborative control hardware device is constructed. This device is used to realize the real-time execution of the power distribution network collaborative control algorithm, the deterministic acquisition of external measurement data, the real-time output of control commands, and the recording and debugging of operating data. The hardware-in-the-loop collaborative control hardware device includes a control calculation and management unit and a main control chip and peripheral interface unit. The two units achieve bidirectional interaction of sampled data, clock synchronization information, control commands, and status information through an internal interconnection channel, thereby meeting the requirements of hardware-in-the-loop closed-loop verification for real-time performance, reliability, and traceability.

[0093] S4: A dynamic display layer 400 that integrates and tracks new energy configuration scheme 102, steady-state indicators and transient response data from multiple sources using a unified time benchmark.

[0094] Specifically, the dynamic display layer 400 is used for visual monitoring and comparative analysis of collaborative planning results and key parameters of power electronic control.

[0095] The dynamic display layer 400 enables visual monitoring and comparative analysis of collaborative planning results and key power electronic control parameters; it constructs a data acquisition, industrial communication and secure transmission, data upload and platform display process for power distribution network collaborative control and hardware-in-the-loop simulation systems.

[0096] It is used to achieve unified aggregation, reliable transmission, traceable storage, and multi-dimensional visualization of data between the mid-level control system and the hardware support side, thereby providing data support for operation monitoring, effect evaluation, and operation and maintenance analysis in scenarios with a high proportion of new energy access.

[0097] The dynamic display layer 400 receives the new energy configuration scheme 102, operating scenarios, and steady-state indicators output by the operation control layer 200. Simultaneously, it receives voltage, frequency, and power electronic control-related data transmitted back from the hardware-in-the-loop verification layer 300, and achieves multi-source data fusion display based on timestamp alignment. The displayed content includes new energy planning scheme information, as well as access node and capacity configuration, energy storage configuration and charging / discharging boundaries, reactive power resource configuration, typical scenarios and disturbance condition numbers, and corresponding steady-state verification results, such as voltage qualification rate, line and transformer current carrying capacity, network loss, and distribution of over-limit locations. The display also includes power electronic control parameters and operating status monitoring based on different power devices, including at least grid connection point voltage and frequency response curves, overshoot, and settling time indicators. It supports curve comparison and indicator summarization under different planning schemes and control parameter versions to form traceable test and evaluation results.

[0098] Furthermore, the dynamic display layer 400 constitutes the system's interactive window, primarily displaying data during system operation. Regarding the result display in the operation control layer 200, it intuitively plots source-load time-series curves and steady-state power flow distribution under various typical scenarios, and presents transient frequency / voltage response waveforms under transient disturbances in real time. The dynamic display layer 400 also supports curve comparison and index summarization under different planning schemes and different control parameter versions, forming traceable test and evaluation results.

[0099] It should be noted that, as Figure 5 As shown, in the dynamic display layer 400, based on the above scheme and control logic, a data acquisition, industrial communication and secure transmission, data upload and platform display process for the distribution network collaborative control and hardware-in-the-loop simulation system is constructed. This is used to realize the unified aggregation, reliable transmission, traceable storage and multi-dimensional visualization of data from the mid-level control system and the hardware support side, thereby providing data support for operation monitoring, effect evaluation and operation and maintenance analysis in the scenario of high proportion of new energy access.

[0100] In this embodiment, the first step is to execute the data output step of the mid-level control system, which generates and outputs equipment status, operating parameters, and fault alarm information. The equipment status includes, but is not limited to, one or more of the following: switch status, power device operating mode, protection activation status, and communication link status. The operating parameters include, but are not limited to, one or more of the following: voltage, current, active power, reactive power, frequency, harmonic indicators, and control command execution quantities. The fault alarm information includes, but is not limited to, one or more of the following: limit exceedance alarms, communication anomaly alarms, execution failure alarms, and safety policy trigger alarms. The above data can be generated by the collaborative control algorithm module, scheduling and strategy module, and operation monitoring module respectively, and output to the data aggregation link according to a preset sampling period or event triggering method.

[0101] The hardware support data acquisition step is then performed to collect and encapsulate hardware-in-the-loop interface access signals and bus data. The hardware support data includes, but is not limited to, one or more of the following: analog or digital sampled values ​​from the HIL interface, serial bus messages, Ethernet messages, and device-side driver feedback information. Variables are identified, timestamps are aligned, and data quality is marked according to the point-to-point mapping relationship to form a data frame that can be used for subsequent unified transmission.

[0102] After data acquisition is completed on the data source side, the industrial communication and secure transmission steps are executed. This is used to uniformly carry and securely transmit data from the mid-level control system and hardware support side via industrial communication protocols. The industrial communication protocols include one or more of MQTT, OPC UA, and Modbus TCP. Secure transmission is implemented using a TLS encrypted channel, thereby ensuring the confidentiality, integrity, and tamper resistance of data during transmission. Furthermore, different quality of service levels and priority policies can be set for different types of data to ensure the real-time delivery of fault alarms and critical operational quantities.

[0103] The data upload and breakpoint resume steps are then executed to send the collected data to the data platform in real-time or scheduled uploads. Local caching and breakpoint resume mechanisms ensure data integrity in the event of network interruption or platform unavailability. Specifically, when an upload failure is detected, the device writes the data to be uploaded to a local cache queue and records the upload offset. After communication is restored, the device continues to upload the unsuccessful portion according to the offset to avoid data loss and duplicate writing. Simultaneously, the cache capacity, retention period, and retry strategy can be configured to adapt to the bandwidth and reliability requirements of different scenarios.

[0104] Finally, the platform performs a multi-dimensional display step to statistically analyze and visualize the uploaded data. The platform provides query and aggregation display based on at least the dimensions of device, feeder, operating condition, and time. It analyzes and displays key indicators including the number of limit violations, adjustment time, overshoot, steady-state error, control output saturation frequency, alarm level distribution, and communication latency statistics. This enables the evaluation of the collaborative control effect under a high proportion of new energy access configurations and the traceability analysis of the hardware-in-the-loop verification process, providing a basis for subsequent strategy optimization and operation and maintenance decisions.

[0105] Example 2, an embodiment of the present invention, provides a distribution network collaborative control and hardware-in-the-loop simulation method for serving high-proportion new energy access configurations, including constructing a source-load probability distribution model 101 based on collected basic data of the distribution network, and generating steady-state scenario sets and transient scenario sets.

[0106] A multi-objective optimization configuration model that takes into account both steady-state and transient indicators is constructed to generate an initial new energy configuration scheme 102. The initial new energy configuration scheme 102 and the scenario set are then sent to the operation control layer 200. A corresponding new energy control module is constructed in the operation control layer, and the control parameters are optimized and iterated based on a reinforcement learning model to ensure good follow of the issued command parameters. The corresponding coupling interface is used to interact with the simulation environment to generate a chained real-time control strategy.

[0107] The real-time control strategy and corresponding control commands are sent to the hardware-in-the-loop verification layer 300 to drive the simulator to run the electromagnetic transient model of the power distribution network, and the response data is collected through the physical interface using hardware coordination devices.

[0108] Based on the feedback dynamic response data, transient safety indicators are output. Combined with the energy management performed by the operation control layer 200, steady-state indicators are obtained and iterated until the optimal configuration scheme set is output.

[0109] Example 3 is an embodiment of the present invention. This embodiment also provides a terminal type hardware device.

[0110] Specifically, the hardware device of the terminal type includes a control, computing and management unit and a main control chip and peripheral interface unit. The two units achieve bidirectional interaction of sampled data, clock synchronization information, control commands and status information through an internal interconnection channel.

[0111] The control operation and management unit includes at least a data acquisition module, a clock synchronization module, a CPU core module, a communication interface module, a power supply and safety module, a local debugging module, an instruction output module, and a log and data storage module. The data acquisition module collects measurements and state variables from the simulation side or the external measured object based on the point table and channel parameters configured in the previous steps, and provides them to the CPU core module after necessary preprocessing and buffering. The clock synchronization module aligns with the external simulation time base or network synchronization signal, generates a unified control cycle trigger signal and timestamp reference, and distributes them to the data acquisition module, CPU core module, instruction output module, and log and data storage module to ensure consistency in sampling, calculation, output, and recording within the same cycle. The CPU core module executes the cooperative control algorithm and real-time task scheduling under the action of the cycle trigger signal, reads the sampling results from the data acquisition module and the external interaction information received by the communication interface module, completes cooperative control calculations, and generates control instructions and constraint boundaries. The communication interface module enables hardware-in-the-loop simulation... The communication connection and protocol adaptation between the real device, host computer, or other collaborative control nodes are completed, including message sending and receiving, encapsulation / decapsulation, caching, and exception handling, and interactive information is provided to the CPU core module. The power supply and safety module is used to stabilize and protect the power supply of the device, and to implement safety enable and exception trigger management for critical modules. When an abnormal condition is detected, an alarm is output and a degradation or safety shutdown strategy is triggered. The local debugging module is used to realize local parameter configuration, online diagnosis, firmware download, and fault location. The instruction output module is used to format the control instructions generated by the CPU core module according to the preset interface format and range / code stream rules, and output them to the simulation object or external execution device to form a closed-loop control. The log and data storage module is used to record and store control process data, event information, alarm information, and operating configuration to support subsequent reproduction analysis and performance evaluation.

[0112] Furthermore, the main control chip and peripheral interface unit are not limited to those mentioned in this invention. In this invention, the main control chip and peripheral interface unit include at least a USB interface, a UART interface, other USB / UART expansion ports, a DDR PHY, an EMMC, a GMAC0, a debug UART, and isolated RS422 and RS485 interfaces, and form a network port and a debug port to the outside. The USB interface, UART interface, and other USB / UART expansion ports are used to achieve general communication connections with host computers, simulation platforms, or maintenance terminals; the DDR PHY is used to connect high-speed volatile memory to carry real-time computing cache and communication queues; the EMMC is used to connect non-volatile storage media to save firmware images, operating configurations, and historical log data; the GMAC0 provides Ethernet communication capabilities through the network port to carry data stream interaction of measurement data, control commands, and synchronization information; the debug UART provides a low-level debugging and online diagnostic channel through the debug port; the isolated RS422 interface and the isolated RS485 interface provide industrial field serial communication capabilities and improve anti-interference performance and system security through electrical isolation to adapt to the differential / bus communication requirements of power distribution automation terminals, protection devices, or simulation interface equipment.

[0113] It should be noted that the closed-loop operation process of this device includes: after the clock synchronization module completes time base alignment and outputs a periodic trigger signal, the data acquisition module completes sampling under the control of the periodic trigger signal and outputs the sampled data to the CPU core module; the CPU core module, in conjunction with the external interactive information received by the communication interface module (including but not limited to simulation-side state variables, operating condition switching commands, and parameter distribution information), performs collaborative control calculations, and the resulting control commands are output to the simulation object or external execution device through the instruction output module, thereby forming a closed-loop link of sampling, calculation, and output; at the same time, the log and data storage module records the control process variables and event information by timestamp and can achieve long-term storage via EMMC; the power supply and safety module continuously monitors the power supply and operating status, and triggers alarms and executes safety controls when an abnormality occurs, so as to ensure the real-time performance, stability, and traceability of the hardware-in-the-loop collaborative control closed-loop operation.

[0114] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A distribution network collaborative control and hardware-in-the-loop simulation system for serving high-proportion renewable energy access configurations, characterized in that, include: The collaborative planning layer (100) is used to construct the source-load probability distribution model (101) and generate a new energy configuration scheme (102) that takes into account both steady-state economic indicators (T1) and transient security indicators (T2) based on the multi-objective optimization algorithm. The operation control layer (200) performs energy management and outputs steady-state indicators in steady state according to the received new energy configuration scheme (102). In transient state, it models different new energy modules based on the characteristics of power electronic devices, and uses reinforcement learning algorithm to perform power electronic parameter self-tuning, outputs transient response data, and issues control commands. The hardware-in-the-loop verification layer (300) is used to build an electromagnetic transient closed-loop simulation environment, receive control commands issued by the operation control layer (200) to perform engineering-level dynamic verification and feed back real-time response data; The dynamic display layer (400) uses a unified time base to display and track the new energy configuration scheme (102), steady-state indicators and transient response data through multi-source fusion.

2. The distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration as described in claim 1, characterized in that: The construction of the source-load probability distribution model (101) includes, Consider the probability distribution model for wind power, photovoltaic power, and load uncertainty; The wind speed frequency distribution characteristics are fitted using a Weibull distribution, and the power characteristic curve of the wind turbine is described using a piecewise function. The Beta distribution is used to describe the probabilistic characteristics of solar irradiance, and the output power of the photovoltaic array is calculated in combination with the ambient temperature. The load power demand characteristics in the distribution network are described using a normal distribution.

3. The distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration as described in claim 1 or 2, characterized in that: The multi-objective optimization algorithm includes, Using the configuration capacity of the components to be planned in the distribution network as the decision variable, the planning solution is performed in the solution space; The configuration scheme and scenario are sent to the operation control module, which then calls simulation calculation or hardware-in-the-loop verification and returns the calculation results. The population is classified by non-dominated sorting, and population diversity is maintained by crowding distance calculation. The Pareto optimal solution set is output as the new energy allocation scheme (102).

4. The distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration as described in claim 3, characterized in that: The operation control layer (200) executes a hierarchical management strategy during steady-state operation to output steady-state economic indicators (T1) and power allocation to new energy sources; In transient response mode, a new energy topology module is constructed, and a transient safety index (T2) is output. The power electronic parameters are self-tuned using a reinforcement learning algorithm.

5. The distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion access configuration of renewable energy as described in claims 1, 2, or 4, characterized in that: The steady-state output steady-state index for energy management under steady-state conditions includes: Under steady-state conditions, energy management and power flow calculations are performed based on the distribution network topology and source-load data. Distributed power sources, energy storage units, and load-side resources are coordinated and optimized for scheduling. Active and reactive power outputs are allocated, and operating costs, power balance constraints, equipment capacity constraints, and voltage constraints are output. Steady-state evaluation indicators such as operating costs, renewable energy absorption rate, voltage qualification rate, and line current carrying rate are also output.

6. The distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration as described in claim 5, characterized in that: The output transient response data includes Under transient conditions, based on the characteristics and application scenarios of different power devices, corresponding new energy topology modules are constructed, and coupling connections between different energy modules are completed. Control commands are generated according to the disturbance scenario and driven to drive the lower-level real-time simulation environment to obtain dynamic response data of frequency and voltage.

7. The distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration as described in claims 1, 2, 4, or 6, characterized in that: The process of receiving control commands issued by the operation control layer (200), performing engineering-level dynamic verification, and feeding back real-time response data includes: Based on steady-state and transient control analysis, an electromagnetic transient closed-loop simulation environment for power distribution networks is constructed under real time scale and combined with the characteristics of real power devices to conduct engineering-level dynamic verification of the cooperative control strategy. Using the RT-LAB real-time simulation platform as the core operating carrier, an electromagnetic transient model including distributed new energy grid connection interface, energy storage unit and load model is built in the real-time simulator. Numerical calculations are performed according to a fixed simulation step size, and node voltages, currents, frequencies, and related state variables are output in real time. The control commands generated by the operation control layer (200) are connected to the RT-LAB system through the physical interface. The RT-LAB updates the model state and feeds back dynamic response data based on the control signals.

8. The distribution network collaborative control and hardware-in-the-loop simulation system for high-proportion renewable energy access configuration as described in claim 7, characterized in that: The dynamic display layer (400) performs visual monitoring and comparative analysis of the collaborative planning results and key parameters of power electronic control; Construct a data acquisition, industrial communication and secure transmission, data upload and platform display process for a power distribution network collaborative control and hardware-in-the-loop simulation system.

9. A distribution network collaborative control and hardware-in-the-loop simulation method for serving a high proportion of renewable energy access configuration, employing the distribution network collaborative control and hardware-in-the-loop simulation system for serving a high proportion of renewable energy access configuration as described in any one of claims 1 to 8, characterized in that, include: Based on the collected basic data of the power distribution network, a source-load probability distribution model (101) is constructed to generate a steady-state scenario set and a transient scenario set; A multi-objective optimization configuration model that takes into account both steady-state and transient indicators is constructed to generate an initial new energy configuration scheme (102). The initial new energy configuration scheme (102) and the scenario set are then sent to the operation control layer (200). A corresponding new energy control module is constructed in the operation control layer, and the control parameters are optimized and iterated based on the reinforcement learning model to ensure good follow of the issued command parameters. It also uses the corresponding coupling interface to interact with the simulation environment and generate a chain-like real-time control strategy; The real-time control strategy and corresponding control commands are sent to the hardware-in-the-loop verification layer (300) to drive the simulator to run the electromagnetic transient model of the power distribution network and collect response data through the physical interface using hardware collaborative devices. Based on the feedback dynamic response data, output transient safety indicators, and combine the energy management of the operation control layer (200) to obtain steady-state indicators and iterate until the optimal configuration scheme set is output.

10. A hardware device of a terminal type, characterized in that, When the computer program is executed by the processor, it implements the distribution network collaborative control and hardware-in-the-loop simulation system for the high proportion of new energy access configuration as described in any one of claims 1 to 8.

Images