Inter-regional potential modeling and optimal control method for industrial adjustable resources in distribution networks

Abstract

This application provides a method for range potential modeling and optimized control of industrial adjustable resources in power distribution networks, relating to the field of power system technology. The method includes: extracting a multi-dimensional resource feature set based on industrial load operation data; calculating potential ranges and generating potential certificates using self-organizing maps; constructing node feature vectors for partitioning and establishing a hierarchical control architecture; generating standardized contract templates and resource priority indices; constructing a control objective function based on the hierarchical control architecture and resource priority indices, with the goal of meeting the predicted future demand of the distribution network and constrained by potential ranges; solving for the optimal control scheme and writing it into a blockchain smart contract; collecting actual execution power; settling and calculating the execution fulfillment rate through the blockchain smart contract; and updating the standardized contract templates and resource priority indices accordingly. This application enables accurate quantification and range representation of the potential of industrial adjustable resources, improving the robustness of control strategies and the reliability of the execution process.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a method for range potential modeling and optimized control of industrial adjustable resources in distribution networks. Background Technology

[0002] With the increasing penetration of new energy sources, the uncertainty and volatility of power distribution network operation have significantly increased. Traditional control methods relying on centralized power sources and peak-shaving units are no longer sufficient to meet the operational requirements that prioritize both safety and economy. Against this backdrop, industrial loads, as important adjustable resources in the power distribution network, are gradually becoming a crucial component of demand response and ancillary services due to their large scale and high flexibility. How to accurately tap the potential of industrial adjustable resources and achieve optimized control has become an important research direction for improving the resilience of power distribution network operation and reducing system operating costs.

[0003] Existing research mainly focuses on two aspects: first, the assessment of adjustable resource potential, which is usually based on static parameter modeling or empirical prediction to quantify the characteristics of industrial load such as interruptibility, transferability, and economy; second, regulation and settlement mechanisms, which mostly adopt centralized optimization, heuristic algorithms, or regional partitioning methods based on electrical distance, combined with static priority ranking strategies to achieve resource scheduling. However, these methods have obvious limitations in practical applications: potential assessment often lacks characterization of uncertainty and dynamism, while contract mechanisms generally rely on predefined templates, making it difficult to adapt to diverse resource characteristics; existing partitioning methods are too simplistic and lack consideration for potential characteristics; optimization models do not explicitly introduce interval constraints, resulting in insufficient robustness; and regulation and settlement rely on centralized platforms, making it difficult to guarantee transparency and credibility. Summary of the Invention

[0004] This application provides a method for range potential modeling and optimized control of adjustable resources in industrial distribution networks to address the problems mentioned in the background art.

[0005] The first aspect of this application provides a method for range potential modeling and optimized control of industrial adjustable resources in distribution networks, including: extracting a multi-dimensional resource feature set based on the operating data of industrial load, classifying the multi-dimensional resource feature set into categories using self-organizing mapping and calculating the corresponding potential ranges, and generating a potential certificate;

[0006] Node feature vectors are constructed and the distribution network is partitioned to establish a hierarchical control architecture. Based on the potential certificate, a standardized contract template is generated and the resource priority index is calculated.

[0007] Based on the hierarchical control architecture and the resource priority index, a control objective function is constructed with the goal of meeting the predicted future demand of the distribution network and the potential range as a constraint. The optimal control scheme is then solved and written into a blockchain smart contract.

[0008] The system collects the actual execution power of the response to the optimal control scheme, executes settlement through the blockchain smart contract and calculates the execution fulfillment rate, and updates the standardized contract template and resource priority index using the actual execution power and the execution fulfillment rate.

[0009] Optionally, in one possible implementation of the first aspect, the operating data of the industrial load includes power curve data and process constraint parameters;

[0010] The multidimensional resource feature set includes interruptibility indicators, transferability indicators, operating cost indicators, process constraint strength indicators, volatility indicators, and mode stability indicators.

[0011] The interruptibility index represents the ratio of allowable energy reduction during a given period to total energy consumption;

[0012] The portability index characterizes the ratio of transferable energy to total energy consumption;

[0013] The operating cost index represents the marginal adjustment cost per unit of adjustment power;

[0014] The process constraint strength index represents the ratio of the locked period to the total period;

[0015] The volatility index characterizes the ratio of the standard deviation to the mean of the power curve;

[0016] The pattern stability index characterizes the entropy value of the daily operating curve cluster distribution.

[0017] Optionally, in one possible implementation of the first aspect, the extraction of a multi-dimensional resource feature set based on industrial load operating data includes:

[0018] Statistical analysis and parameter extraction are performed on the power curve data and the process constraint parameters to obtain the overall characteristics of the industrial load. The overall characteristics include power mean parameters, power discrete parameters, energy efficiency index, and response failure rate.

[0019] The power curve data is divided into multiple process stage intervals based on the process constraint parameters, and the stage characteristics of each process stage interval are collected. The stage characteristics include the power mean parameter, power discrete parameter, minimum process start and stop time, process stage label, and process constraint parameter set within the stage.

[0020] The collected overall features and stage features are integrated to obtain a basic feature set; the multidimensional resource feature set is calculated based on the basic feature set.

[0021] Optionally, in one possible implementation of the first aspect, the step of classifying the multidimensional resource feature set using a self-organizing map and calculating the corresponding potential interval includes:

[0022] Construct a self-organizing map two-dimensional grid network, calculate the Euclidean distance between the multidimensional resource feature set and each node in the self-organizing map two-dimensional grid network, and determine the node with the smallest distance as the best matching node;

[0023] The multidimensional resource feature set is mapped to the best matching node to form a category set;

[0024] For each category in the category set, power samples within the allowable adjustment period are extracted from the operational data, and the upper and lower limits of the potential for that category are calculated using the quantile method to obtain the potential range;

[0025] The energy consumption ratio of a single industrial adjustable resource is calculated based on operational data, and the potential range is allocated to the single industrial adjustable resource according to the energy consumption ratio.

[0026] Optionally, in one possible implementation of the first aspect, the construction of node feature vectors and partitioning of the distribution network to establish a hierarchical control architecture includes:

[0027] Obtain the node electrical parameters of the distribution network, including the node equivalent impedance, voltage response parameter, and node load power parameter;

[0028] The average adjustable potential and potential uncertainty parameters of resources within a node are calculated based on potential certificates.

[0029] A comprehensive feature vector is constructed and normalization is performed to obtain a node feature matrix. The comprehensive feature vector includes node electrical parameters, average adjustable potential, and potential uncertainty parameters.

[0030] Perform two-dimensional clustering on the node feature matrix, calculate the distance from the node to the cluster center in the electrical dimension and the distance in the potential dimension, and determine the comprehensive distance. Based on the comprehensive distance, the node is assigned to the corresponding region control unit.

[0031] The hierarchical control architecture is established based on regional control units; the regional control units include a scheduling center level, a regional control level, and a terminal resource level.

[0032] Optionally, in one possible implementation of the first aspect, generating a standardized contract template based on the potential certificate includes:

[0033] Extract the potential range, the mean potential parameter, and the potential uncertainty parameter from the potential certificate;

[0034] Obtain the preset adjustment cost function and response time window parameters;

[0035] The standardized contract template is generated using the potential range, the potential mean parameter, the potential uncertainty parameter, the adjustment cost function, and the response time window parameter.

[0036] The adjustment cost function is constructed based on the baseline cost and slope coefficient; the response time window parameters include the allowable response delay time and duration.

[0037] Optionally, in one possible implementation of the first aspect, the computing resource priority index includes:

[0038] Extract potential level indicators and reliability indicators from potential certificates, and extract response speed indicators and economic indicators from standardized contract templates;

[0039] The potential level index, the reliability index, the response speed index, and the economic index are standardized.

[0040] Based on a preset proportional parameter vector, a weighted summation is performed on each standardized indicator to obtain the initial priority index;

[0041] The risk correction parameter is calculated based on the potential range, and the initial priority index is corrected using the risk correction parameter to obtain the resource priority index.

[0042] The risk correction parameter is used to reduce the priority of resources whose fluctuation range in the potential range exceeds a preset range.

[0043] Optionally, in one possible implementation of the first aspect, the construction of the control objective function that aims to meet the predicted future demand of the distribution network, is based on the resource priority index and constrained by the potential range, specifically involves establishing a control objective function that takes into account the system load tracking accuracy, operating costs, and risk costs.

[0044] The system load tracking accuracy characterizes the degree of matching between the adjusted power and the future demand;

[0045] The operating cost is calculated based on the adjusted cost function in the standardized contract template; the risk cost is a correction term for deviation from the potential range;

[0046] The potential range is used as a constraint condition for the control objective function, and response time constraints, node voltage constraints, and regional power balance constraints are set.

[0047] Optionally, in one possible implementation of the first aspect, the execution of settlement via the blockchain smart contract includes:

[0048] Call standardized contract templates via blockchain smart contracts;

[0049] Calculate the deviation between the actual power output and the planned power output indicated by the optimal control scheme;

[0050] Based on the adjustment cost function in the standardized contract template and the deviation value, calculate the settlement interaction value;

[0051] The execution fulfillment rate is calculated based on the ratio of the deviation value to the planned power.

[0052] Optionally, in one possible implementation of the first aspect, updating the standardized contract template and resource priority index using the actual execution power and the execution fulfillment rate includes:

[0053] The mean deviation and the deviation fluctuation are calculated based on the deviation value, and the mean deviation and the deviation fluctuation are combined to form the execution error parameter;

[0054] The potential uncertainty parameters in the standardized contract template are corrected using the execution error parameters.

[0055] Based on the preset memory decay parameter, the resource priority index is iteratively updated in combination with the execution fulfillment rate.

[0056] The beneficial effects of this application are as follows:

[0057] 1. This application realizes the refined and intervalized characterization of industrial adjustable resource potential. By extracting a multi-dimensional resource feature set based on operational data, and using self-organizing mapping and quantile method to calculate potential intervals and generate potential certificates, this application not only breaks through the information barrier of traditional single-dimensional load modeling, but also transforms the originally vague load curve into a digital profile with clear upper and lower limits. This endows the distribution network with the microscopic perception capability of resource interruptibility, transferability and operational costs, and provides standardized data support for precise scheduling.

[0058] 2. This application directly writes standardized contract templates and optimal control schemes into blockchain smart contracts, using smart contracts to automatically collect actual execution power and calculate the execution fulfillment rate. This automated execution method based on on-chain code establishes an immutable trust anchor between supply and demand, eliminates opaque links in the traditional scheduling and execution process, and ensures that every power adjustment instruction has extremely high traceability and execution certainty, thereby creating a highly transparent and automatically operating trusted control environment.

[0059] 3. This application constructs a control objective function that includes potential range constraints and a resource priority index, and dynamically adjusts the priority by combining risk correction parameters. This effectively isolates the operational risks caused by resource heterogeneity. Simultaneously, by using the performance rate to iteratively update the contract template parameters and priority index, the control system can self-evolve in a dynamically changing power grid environment, achieving a balance between safety and economy, and enhancing the distribution network's resilience in the face of source-load fluctuations. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating the method for range potential modeling and optimized control of adjustable resources in the distribution network industry provided in this application embodiment;

[0061] Figure 2 This is a schematic diagram of a data-driven interval potential partitioning and evaluation mechanism for an interval potential modeling and optimized control method for adjustable resources in the distribution network industry provided in this application embodiment;

[0062] Figure 3 This is a schematic diagram of a two-dimensional partitioning method for the interval potential modeling and optimized control method of adjustable resources in the distribution network industry provided in the embodiments of this application;

[0063] Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0065] The technical solutions of this application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0066] See Figure 1 This is a flowchart illustrating the method for range potential modeling and optimized control of adjustable resources in industrial distribution networks provided in this application. Figure 1The execution entity of the method shown can be a software and / or hardware device. The execution entity of this application can include, but is not limited to, at least one of the following: user equipment, network equipment, etc. User equipment can include, but is not limited to, computers, smartphones, personal digital assistants (PDAs), and the aforementioned electronic devices. Network equipment can include, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud based on cloud computing consisting of a large number of computers or network servers. Cloud computing is a type of distributed computing, consisting of a super virtual computer composed of a group of loosely coupled computers. This embodiment does not limit this. Steps 100 to 400 are detailed as follows:

[0067] Step 100: Extract a multidimensional resource feature set based on the industrial load operation data, classify the multidimensional resource feature set into categories using self-organizing mapping and calculate the corresponding potential intervals, and generate a potential certificate.

[0068] Specifically, the operating data for industrial loads includes power curve data and process constraint parameters.

[0069] In some embodiments, the "extraction of multi-dimensional resource feature set based on industrial load operation data" in step 100 specifically involves collecting the average power parameter, discrete power parameter, minimum start-up and shutdown time of the process, process stage label, set of process constraint parameters, energy efficiency index, and response failure rate of the industrial load, and integrating the collected overall features with the stage features.

[0070] Specifically, statistical analysis and parameter extraction are performed on the power curve data and process constraint parameters to obtain the overall characteristics of the industrial load.

[0071] The overall characteristics include power mean parameters, power discrete parameters, energy efficiency indicators, and response failure rate.

[0072] The power curve data is divided into multiple process stage intervals based on process constraint parameters, and the stage characteristics of each process stage interval are collected.

[0073] The stage characteristics include the average power parameter, power discrete parameter, minimum process start-up and shutdown time, process stage label, and set of process constraint parameters within the stage.

[0074] The collected overall features and stage features are integrated to obtain a basic feature set; a multidimensional resource feature set is then calculated based on the basic feature set.

[0075] For example, to address the issues of diverse industrial load types, complex process constraints, and heterogeneous data sources in distribution networks, a fusion modeling and unified extraction method is proposed. By constructing a process-load joint feature model, the traditional method relies solely on load curves while ignoring process constraints, thus achieving a complete expression of multi-dimensional adjustable attributes and providing a solid data foundation for subsequent potential assessment and contract construction.

[0076] Collect industrial operation data from different levels:

[0077] (1) Electrical side data: Active power curves provided by smart meters The sampling period is , i represents the index number of the i-th industrial load, i=1,2,...,N, where N is the total number of industrial loads; t represents the discrete-time index; m indicates that the data is collected through a smart meter;

[0078] (2) Operation monitoring data: Equipment start-up and shutdown status collected by the SCADA system With cumulative runtime The sampling period is ;in, Indicates the first The industrial load at any time The device's operating status is indicated by a value of 1, which represents the device being in operation, and a value of 0, which represents the device being stopped. Indicates the first Industrial load cutoff time Cumulative runtime; This indicates the sampling time interval of the SCADA system; s indicates that the data is the operating status data collected by the SCADA system.

[0079] (3) Process production data: Production stage labels provided by the manufacturing execution system. With process constraint parameters ,in, Minimum start-stop time, Due to production constraints, Energy efficiency factor.

[0080] To eliminate the sampling frequency differences between data from different sources, linear interpolation and moving average methods were used to unify all data to the same time resolution. To create standardized multi-source time-series datasets:

[0081] ;

[0082] Based on a standardized dataset, a fusion feature reflecting the electrical characteristics and process constraints of industrial loads is constructed. For industrial loads... Its eigenvectors are represented as:

[0083] ;

[0084] in, The average power is represented by T, which represents the total number of sampling periods within the statistical period. Indicates the standard deviation of power fluctuation. , These represent the minimum start-up and shutdown times of the process. Labels indicating process stages Represents the set of process constraint parameters; Indicates energy efficiency indicators; This indicates the historical response failure rate, among which, These represent the total number of failed data and the total number of data, respectively, reflecting the reliability level.

[0085] To ensure the physical interpretability of the features, a process tagging mechanism is introduced. For each power curve, based on... It is divided into multiple process stage intervals:

[0086] ;

[0087] in, Indicates the first The start and end times of an industrial load in the first process interval; Indicates the first The total number of process stages per industrial load; The label number represents the current process stage, and its value range is [value range missing]. to .

[0088] Within each stage, average power, fluctuation coefficient, and constraint parameters are extracted to obtain a staged feature vector:

[0089] in, These represent the i-th industrial load in the i-th... Average power and standard deviation of power fluctuation within each process stage; Indicates the first Each process stage; , These represent the minimum start-up and shutdown times corresponding to each process stage; , These represent the production constraints and energy efficiency factors corresponding to this process stage, respectively.

[0090] In this way, each resource not only has overall characteristics It also has phased characteristics. This is more suitable for subsequent potential range assessment.

[0091] By integrating the overall characteristics and stage characteristics of all resources, a global feature set is obtained:

[0092] ;

[0093] Where F represents the global feature set formed by integrating the overall and stage characteristics of all industrial loads; N represents the total number of industrial loads; Indicates the first The overall characteristic vector of an industrial load; Indicates the first A staged feature vector of an industrial load at each process stage.

[0094] Z-score normalization is applied to the feature set:

[0095] ;

[0096] in, This represents the original value of resource i in feature dimension j. feature dimensions The mean and standard deviation. This represents the standardized feature values. Std indicates standardization. The standardized feature set obtained through Z-score standardization not only facilitates clustering and modeling but also ensures the comparability of different feature dimensions.

[0097] To address the problems in assessing the adjustable resource potential of industry, such as the reliance on single indicators, strong subjectivity in clustering, and a lack of uncertainty in potential values, this application constructs a multi-dimensional indicator system, employs self-organizing mapping clustering, and uses interval estimation methods to achieve zonal quantification and interval representation of resource potential. This provides standardized input for subsequent contract design and regulation optimization. Figure 2 As shown.

[0098] It should be noted that the multidimensional resource feature set includes interruptibility indicators, transferability indicators, operating cost indicators, process constraint strength indicators, volatility indicators, and mode stability indicators.

[0099] Among them, the interruptibility index represents the ratio of allowable energy reduction to total energy consumption; the transferability index represents the ratio of transferable energy to total energy consumption; the operating cost index represents the marginal adjustment cost per unit of adjustable power; the process constraint strength index represents the ratio of locked-in time period to total time period; the volatility index represents the ratio of the standard deviation to the mean of the power curve; and the model stability index represents the entropy value of the cluster distribution of the daily operating curve.

[0100] To address the issue of insufficient dimensions in expressing resource characteristics, six categories of indicators (i.e., interruptibility indicators, transferability indicators, operating cost indicators, process constraint strength indicators, volatility indicators, and mode stability indicators) are selected, covering three aspects: adjustability, economic characteristics, and operating modes. For each industrial adjustable resource... Define the indicator vector:

[0101] ;

[0102] in, The interruptibility index is defined as the period during which energy can be reduced. Total energy consumption The ratio; The transferability index is defined as the energy that can be translated. The ratio of total energy consumption; The operating cost indicator is defined as the unit regulation power. marginal cost ; The process constraint strength index is defined as the lock-in period. Total time period The ratio; The volatility index is defined as the ratio of the standard deviation to the mean of the power curve. The stability index of the pattern is defined as the cluster distribution of the daily running curves. The entropy value.

[0103] The six categories of indicators were standardized using Z-score:

[0104] ;

[0105] in, This represents the original index vector of industrial adjustable resources r in feature dimension j. , The first The mean and standard deviation of the original indicator vector; This represents the standardized index value of the r-th industrial adjustable resource on feature dimension j.

[0106] Furthermore, after standardization, the standardized index matrix M is obtained:

[0107] ;

[0108] in, This is the standardized index vector; R represents the total number of industrial adjustable resources.

[0109] In some embodiments, step 100, "classifying the multidimensional resource feature set using a self-organizing map and calculating the corresponding potential interval," specifically includes steps 110 to 130:

[0110] Step 110: Construct a self-organizing map two-dimensional grid network, calculate the Euclidean distance between the multidimensional resource feature set and each node in the self-organizing map two-dimensional grid network, and determine the node with the smallest distance as the best matching node.

[0111] Step 120: Map the multidimensional resource feature set to the best matching node to form a category set.

[0112] Step 130: For each category in the category set, extract the power samples that are in the allowable adjustment period from the running data, and use the quantile method to calculate the upper and lower limits of the category's potential to obtain the potential range.

[0113] Step 140: Calculate the energy consumption ratio of a single industrial adjustable resource based on the operating data, and allocate the potential range to the single industrial adjustable resource according to the energy consumption ratio.

[0114] To eliminate the subjectivity of manually specifying the number and boundaries of categories, a self-organizing map (SOM) is used for adaptive clustering. A two-dimensional grid network is constructed. ,in, Let be the weight vector of the j-th node, with a dimension of 6. N represents a two-dimensional grid network of self-organizing maps; j represents the index number of a node in the network; then, find the distance input index vector. The nearest node :

[0115] ;

[0116] in, Let be the standardized index vector of the r-th resource; This represents the Euclidean distance.

[0117] For the nearest node and its neighboring nodes Perform weight updates:

[0118] ;

[0119] in, Indicates the first The node at the th The weight vector at the next iteration; This represents the updated weight vector; For the first Learning rate at the next iteration; For the first The neighborhood radius at the next iteration; when the quantization error converges or the number of iterations reaches the upper limit. Training should be stopped at this time. Represents the nearest node With neighboring nodes The topological distance between them on the grid. All resources are mapped to their best matching cell node, i.e., the nearest node. Forming a category set ,in Indicates the first The set of industrially adjustable resources included in each category. This is the category index number.

[0120] To address the issue of traditional point-value potential assessment neglecting uncertainty, this application introduces interval-based potential representation. For the category set... Collect all member resources during the permitted adjustment period. Power samples:

[0121] ;

[0122] in, This represents the set of power samples for the k-th category; Let represent the power value of the r-th industrial adjustable resource at time t.

[0123] The upper and lower limits of category potential are calculated using the quantile method to obtain the estimation interval. :

[0124] ;

[0125] ;

[0126] in, These represent the lower quantile and the upper quantile, respectively. This represents the quantile coefficient, which is 0.2 in this application; This represents the lower bound of the potential for the k-th category. This represents the upper limit of the potential for the k-th category.

[0127] For a single resource r within the interval, based on its energy consumption percentage distribute:

[0128] ;

[0129] in, This represents the potential range of the r-th industrial adjustable resource; This represents the energy consumption weight of the r-th resource within its category k.

[0130] In this way, the adjustable potential of each resource is expanded from a single value to an upper and lower limit range.

[0131] To facilitate subsequent scheduling and contract mechanisms, resources are categorized into different levels. The average resource potential is defined as follows:

[0132] ;

[0133] in, This represents the average potential of the r-th industrial adjustable resource; This represents the lower limit of the r-th industrial adjustable resource potential range; This represents the upper limit of the r-th industrial adjustable resource potential range.

[0134] Set threshold set ,in, , , The preset potential level classification threshold is used, and the following conditions are met: Resources are divided into four categories:

[0135] Low potential;

[0136] Medium potential;

[0137] High potential;

[0138] Extremely high potential;

[0139] Simultaneously define the uncertainty coefficient:

[0140] ;

[0141] in, Indicates the first The uncertainty coefficient of an industrial adjustable resource is used to measure the relative volatility of the resource's potential range. Indicates the first The upper limit of the adjustable resource potential range for an industry; Indicates the first The lower limit of the range of adjustable resource potential for an industry.

[0142] Furthermore, to ensure the results can be directly used subsequently, the interval, level, and uncertainty parameters are encapsulated into potential certificates and bound to the partitioning results. A potential certificate is generated for each industrial adjustable resource r:

[0143] ;

[0144] in, Let r be the grade label for the r-th industrial adjustable resource. Let be the average potential of the r-th industrial adjustable resource. Then, let the potential of the r-th industrial adjustable resource be... Clustering of resources Binding, forming interval potential labels:

[0145] ;

[0146] in, This represents the interval potential label for the r-th industrial adjustable resource.

[0147] All Unified storage in potential database It serves as the sole input interface for the contract mechanism and priority control in subsequent steps.

[0148] Step 200: Construct node feature vectors and partition the distribution network, establish a hierarchical control architecture, generate standardized contract templates based on the potential certificates, and calculate the resource priority index.

[0149] To address the issues of inflexible matching methods and insufficient contract execution credibility in the process of regulating industrial adjustable resources, this application's embodiments form a complete closed-loop regulation and settlement mechanism through standardized self-executing contract template modeling, prediction-driven dynamic matching mechanism, on-chain trusted execution of contracts, and result feedback iterative optimization, ensuring matching effectiveness and contract credibility under dynamic changes in supply and demand.

[0150] In some embodiments, step 200 includes steps 210 to 230:

[0151] Step 210: Construct node feature vectors and partition the distribution network to establish a hierarchical control architecture.

[0152] Specifically, step 210 includes steps 211 to 215:

[0153] Step 211: Obtain the node electrical parameters of the distribution network, including the node equivalent impedance, voltage response parameters, and node load power parameters.

[0154] The electrical parameters of each node are obtained from the distribution network topology data and power flow calculation results. The equivalent impedance of a node is extracted from the corresponding diagonal elements of the distribution network impedance matrix, reflecting the electrical distance between the node and the system power source. Voltage response parameters include active voltage response parameters and reactive voltage response parameters. The active voltage response parameter characterizes the node voltage change caused by a unit change in active power, and the reactive voltage response parameter characterizes the node voltage change caused by a unit change in reactive power. Both active and reactive voltage response parameters are obtained from the corresponding elements of the sensitivity matrix obtained by inverting the Jacobian matrix. Node load power parameters include active load power parameters and reactive load power parameters, which characterize the active and reactive power consumed by the node, respectively.

[0155] The node electrical parameter vector is formed by combining the node equivalent impedance, active voltage response parameter, reactive voltage response parameter, active load power parameter, and reactive load power parameter.

[0156] Step 212: Calculate the average adjustable potential and potential uncertainty parameters of resources within the node based on the potential certificate.

[0157] In a power distribution network, a single node often connects to multiple industrial adjustable resources. The potential certificate for each industrial adjustable resource records its own potential mean and uncertainty coefficient. To facilitate zoning and control at the node level, it is necessary to aggregate the potential information of all industrial adjustable resources within the node.

[0158] The potential certificates of all industrial adjustable resources within a node are read from the potential database, and the mean potential and uncertainty coefficient of each industrial adjustable resource are extracted. The arithmetic mean of the mean potentials of all industrial adjustable resources within a node is calculated to obtain the node's average adjustable potential. The arithmetic mean of the uncertainty coefficients of all industrial adjustable resources within a node is calculated to obtain the node's potential uncertainty parameter. The average adjustable potential characterizes the overall adjustment capability of the node, while the node's potential uncertainty parameter characterizes the stability of the node's adjustment capability.

[0159] Step 213: Construct a comprehensive feature vector and perform normalization to obtain the node feature matrix. The comprehensive feature vector includes node electrical parameters, average adjustable potential, and potential uncertainty parameters.

[0160] Traditional distribution network zoning methods only consider electrical characteristics, dividing nodes into different regions based on electrical distance or voltage sensitivity. This method ignores the differences in the potential distribution of adjustable resources within a node, which may lead to significant disparities in resource potential within the same region, making it difficult for regional control units to coordinate scheduling. Therefore, this application considers both electrical parameters and potential parameters together to construct a comprehensive feature vector.

[0161] The node electrical parameter vector, average adjustable potential, and node potential uncertainty parameter are concatenated to form the node's comprehensive feature vector. The comprehensive feature vector contains seven components: the first five components come from the node electrical parameter vector, and the last two components are the average adjustable potential and the node potential uncertainty parameter, respectively.

[0162] Because the dimensions and numerical ranges of the nodal electrical parameters, average adjustable potential, and nodal potential uncertainty parameters differ, normalization processing is required for the composite eigenvector. For each component of the composite eigenvector, the maximum and minimum values ​​of that component across all nodes are calculated. The value of each node for that component is then subtracted from the minimum value and divided by the difference between the maximum and minimum values, ensuring that the normalized component value falls within the range of 0 to 1. The normalized composite eigenvectors of all nodes are then arranged in rows to form the nodal feature matrix.

[0163] Step 214: Perform two-dimensional clustering on the node feature matrix, calculate the distance from the node to the cluster center in the electrical dimension and the distance in the potential dimension, and determine the comprehensive distance. Based on the comprehensive distance, assign the node to the corresponding regional control unit.

[0164] Specifically, single-dimensional clustering cannot adequately account for both electrical and potential characteristics. If clustering is based solely on electrical characteristics, nodes with vastly different potentials may be grouped into the same region, leading to difficulties in coordinating regulation within that region. Conversely, if clustering is based solely on potential characteristics, nodes that are electrically distant may be grouped into the same region, resulting in increased delays in the transmission of regulation commands and an increased risk of voltage exceeding limits.

[0165] The node characteristic matrix is ​​divided into an electrical dimension submatrix and a potential dimension submatrix. The electrical dimension submatrix contains five columns corresponding to the node equivalent impedance, active voltage response parameter, reactive voltage response parameter, active load power parameter, and reactive load power parameter. The potential dimension submatrix contains two columns corresponding to the average adjustable potential and the node potential uncertainty parameter.

[0166] Clustering is performed on the electrical dimension submatrix and the potential dimension submatrix respectively using the K-means method to obtain the set of cluster centers for the electrical dimension and the set of cluster centers for the potential dimension. The number of cluster centers is equal to the preset number of regional control units.

[0167] For each node, calculate the Euclidean distance from the node's corresponding row vector in the electrical dimension submatrix to each electrical dimension cluster center, and the Euclidean distance from the node's corresponding row vector in the potential dimension submatrix to each potential dimension cluster center. For the same cluster center number, perform a weighted sum of the electrical dimension Euclidean distances and potential dimension Euclidean distances based on a balancing parameter to obtain the comprehensive distance from the node to that cluster center. The balancing parameter ranges from 0 to 1; a larger balancing parameter indicates a higher importance of the electrical dimension in the partition, while a smaller balancing parameter indicates a higher importance of the potential dimension in the partition. The balancing parameter, ranging from 0 to 1, is used to adjust the relative importance of the electrical and potential dimensions in the partition.

[0168] Traverse all cluster center numbers, determine the cluster center number with the smallest comprehensive distance, and assign the node to the region control unit corresponding to that number.

[0169] Step 215: Establish a hierarchical control architecture based on the regional control unit; the regional control unit includes the scheduling center level, the regional control level, and the terminal resource level.

[0170] like Figure 3 As shown, this application establishes a three-tiered hierarchical control architecture based on the regional control unit.

[0171] The dispatch center is located at the top level of the hierarchical control architecture. It is responsible for receiving information on the operation status of the distribution network and load forecast data, formulating global control targets, and decomposing the control targets into regional control instructions for each regional control unit and issuing them to the regional control level.

[0172] The regional control unit is located in the middle layer of the hierarchical control architecture, and each regional control unit corresponds to a regional controller. After receiving the regional control instructions issued by the dispatch center, the regional controller further decomposes the regional control instructions into execution instructions for each industrial adjustable resource based on the potential certificates and standardized contract templates of each node in the region, and issues them to the terminal resource level.

[0173] The terminal resource level is located at the bottom of the hierarchical control architecture and includes all industrial adjustable resources. Each industrial adjustable resource receives execution instructions from the regional control level, performs adjustment actions, and feeds back the actual execution power and execution status to the regional control level. The regional control unit aggregates the feedback information from all industrial adjustable resources within its region and reports it to the dispatch center level, forming a closed loop between the issuance of control instructions and execution feedback.

[0174] Step 220: Generate a standardized contract template based on the potential certificate.

[0175] Specifically, step 220 includes steps 221 to 223:

[0176] Step 221: Extract the potential range, the potential mean parameter, and the potential uncertainty parameter from the potential certificate.

[0177] The potential certificates of industrial adjustable resources are retrieved from the potential database, and the potential range is extracted. The potential range consists of a lower potential limit and an upper potential limit. The lower potential limit represents the minimum adjustable power of industrial adjustable resources under conservative estimation conditions, and the upper potential limit represents the maximum adjustable power of industrial adjustable resources under optimistic estimation conditions.

[0178] The potential mean parameter characterizes the expected adjustable power of industrial adjustable resources.

[0179] The potential uncertainty parameter is the ratio of the difference between the upper and lower potential limits to the sum of the upper and lower potential limits, characterizing the relative fluctuation range of the potential range of industrial adjustable resources. A larger value for the potential uncertainty parameter indicates a wider range of variation in the actual adjustable capacity of industrial resources, and a lower certainty in the control outcome.

[0180] Step 222: Obtain the preset adjustment cost function and response time window parameters.

[0181] The preset adjustment cost function and response time window parameters are obtained from the operating parameter database of industrial adjustable resources.

[0182] The adjustment cost function is used to calculate the adjustment cost of industrial adjustable resources at a given adjustment power. Different types of industrial adjustable resources have different cost structures; metallurgical furnaces have higher adjustment costs due to high requirements for process continuity, while air conditioning loads have lower adjustment costs due to their high flexibility. The adjustment cost function is constructed based on the baseline cost and the slope coefficient. The baseline cost represents the fixed cost before adjustment, and the slope coefficient represents the additional marginal cost required for each additional unit of adjustment power. The adjustment cost function multiplies the adjustment power by the slope coefficient and adds the baseline cost to obtain the total adjustment cost.

[0183] The response time window parameter includes the allowable response delay time and the duration. The allowable response delay time represents the maximum permissible time interval between receiving a control command and starting to execute a response from the industrial adjustable resource. Some industrial loads cannot respond to control commands immediately due to process constraints and require reserved start-up and shutdown preparation time. The duration represents the length of time that the industrial adjustable resource can maintain a regulated state, which is limited by equipment heat capacity and process cycle time.

[0184] Step 223: Generate a standardized contract template using the potential range, the potential mean parameter, the potential uncertainty parameter, the adjustment cost function, and the response time window parameter.

[0185] The adjustment cost function is constructed based on the baseline cost and slope coefficient; the response time window parameters include the allowable response delay time and duration.

[0186] By combining and encapsulating the potential range, the mean potential parameter, the potential uncertainty parameter, the adjustment cost function, and the response time window parameter, a standardized contract template for industrial adjustable resources is formed.

[0187] The standardized contract template uses a unified data structure for storage, containing five fields: the potential mean parameter field records the expected adjustable power of industrial adjustable resources; the potential range field records the upper and lower limits of the adjustable power of industrial adjustable resources; the potential uncertainty parameter field records the relative fluctuation range of the potential range of industrial adjustable resources; the adjustment cost function field records the benchmark cost and slope coefficient, which are used to calculate the adjustment cost during subsequent settlement; and the response time window parameter field records the allowed response delay time and duration, which are used to verify the timing constraints when generating subsequent control schemes.

[0188] To ensure the tamper-proof and traceability of standardized contract templates, a hash digest of the template content is calculated and combined with the digital signature of the industrial adjustable resources, then written into the blockchain to form a unique on-chain index. During subsequent implementation and settlement of regulatory schemes, the standardized contract template is retrieved through this unique on-chain index, ensuring that the contract parameters used by both supply and demand parties are consistent and irreversible.

[0189] For example, the potential certificate output based on the aforementioned steps is a resource. Generate standardized contract templates :

[0190] ;

[0191] This standardized contract template includes a potential range, mean, uncertainty parameter, price function, and response time window. Its structure is defined as follows:

[0192] ;

[0193] in, Indicates the average potential; Indicates the potential of the interval; Indicates the uncertainty coefficient; For the quotation function, and These are the benchmark price and the slope coefficient, respectively. This represents the response time parameter, where For the allowable response delay, The duration is specified. All contract templates are stored in a structured format and uniquely indexed by being written on-chain with hash digests and digital signatures.

[0194] Step 230: Calculate the resource priority index.

[0195] Specifically, step 230 includes steps 231 to 234:

[0196] Step 231: Extract potential level indicators and reliability indicators from the potential certificate, and extract response speed indicators and economic indicators from the standardized contract template.

[0197] The resource priority index is used to rank various adjustable resources in industry when selecting control resources. Traditional methods only rank resources according to their adjustable power, ignoring differences in resource reliability, response speed, and economic efficiency. This may result in priority resources having high potential but low performance rates or high adjustment costs. Therefore, this application comprehensively considers multiple dimensions of indicators to calculate the resource priority index.

[0198] Potential level indicators and reliability indicators are extracted from the potential certificate. The potential level indicator is determined by comparing the average potential parameter with a preset threshold; the higher the average potential parameter, the higher the potential level indicator. The reliability indicator is calculated based on the historical performance rate; the higher the historical performance rate, the higher the reliability indicator.

[0199] Response speed and economic indicators are extracted from standardized contract templates. The response speed indicator is calculated based on the allowable response delay time; the shorter the allowable response delay time, the higher the response speed indicator. The economic indicator is calculated based on the slope coefficient in the regulation cost function; the smaller the slope coefficient, the lower the cost per unit of regulation power, and the higher the economic indicator.

[0200] Step 232: Standardize the potential level indicators, reliability indicators, response speed indicators, and economic indicators.

[0201] Because the potential level, reliability, response speed, and economic indicators have different dimensions and numerical ranges, directly weighting and summing them would lead to the indicators with larger numerical ranges dominating the priority results. Therefore, it is necessary to perform standardization processing on each of the four indicators separately.

[0202] For each indicator, calculate the mean and standard deviation of the indicator across all industrial adjustable resources. Subtract the mean from the value of each industrial adjustable resource for the indicator and divide by the standard deviation to make the standardized indicator value follow a distribution with a mean of zero and a standard deviation of 1.

[0203] Step 233: Perform a weighted summation on each standardized index based on the preset proportional parameter vector to obtain the initial priority index.

[0204] The preset proportional parameter vector contains four components, corresponding to the weighted proportions of potential level indicators, reliability indicators, response speed indicators, and economic indicators, respectively. The sum of the four weighted proportions is 1, and the magnitude of each weighted proportion reflects the degree of emphasis the dispatch center places on different evaluation dimensions. When emphasizing adjustment capability, the weighted proportion of the potential level indicator is increased; when emphasizing execution determinism, the weighted proportion of the reliability indicator is increased; when emphasizing rapid response, the weighted proportion of the response speed indicator is increased; and when emphasizing economic efficiency, the weighted proportion of the economic indicator is increased.

[0205] Multiply the standardized potential level index by the corresponding weighting ratio, calculate the product of the standardized reliability index and the corresponding weighting ratio, calculate the product of the standardized response speed index and the corresponding weighting ratio, calculate the standardized economic index by the corresponding weighting ratio, and add the four products together to obtain the initial priority index of industrial adjustable resources.

[0206] Step 234: Calculate the risk correction parameter based on the potential range, and use the risk correction parameter to correct the initial priority index to obtain the resource priority index.

[0207] Among them, the risk correction parameter is used to reduce the priority of resources whose volatility in the potential range exceeds the preset range.

[0208] The initial priority index only reflects the overall performance of industrial adjustable resources across various evaluation dimensions, without considering the regulatory risks arising from fluctuations in the potential range. If the fluctuation range of the potential range of industrial adjustable resources is too large, the actual adjustable power may deviate significantly from the expected adjustable power, resulting in the regulatory scheme's implementation effect falling short of expectations.

[0209] The risk adjustment parameter is calculated based on the potential range. The difference between the upper and lower potential limits is divided by the potential mean parameter to obtain the relative volatility. This relative volatility is compared to a preset risk threshold. If the relative volatility exceeds the preset risk threshold, the risk adjustment parameter is less than 1, reducing the initial priority index. If the relative volatility does not exceed the preset risk threshold, the risk adjustment parameter is equal to 1, and the initial priority index remains unchanged. The greater the relative volatility exceeds the preset risk threshold, the smaller the risk adjustment parameter value, and the greater the penalty to the initial priority index.

[0210] The resource priority index is obtained by multiplying the initial priority index by the risk correction parameter. The higher the resource priority index, the higher the ranking of industrial adjustable resources in the selection of control resources, and the greater the likelihood of them being selected first.

[0211] Step 300: Based on the hierarchical control architecture and the resource priority index, construct a control objective function that aims to meet the predicted future demand of the distribution network and is constrained by the potential range. Solve the function to generate the optimal control scheme and write it into the blockchain smart contract.

[0212] To address the issues of uneven resource distribution, inconsistent control granularity, and uncertainty of potential affecting regional control in power distribution networks, this application establishes a hierarchical control model by combining electrical characteristics and potential certificates to achieve hierarchical management of dispatch centers, regional control units, and terminal resources.

[0213] In some embodiments, step 300 includes steps 310 to 330:

[0214] Step 310: Obtain the future demand of the distribution network and determine the target regulation power for each period within the regulation cycle.

[0215] Specifically, step 310 includes steps 311 to 313:

[0216] Step 311: Obtain the predicted load power of the distribution network for each period within the control cycle from the load forecasting module.

[0217] The dispatch center receives future demand data for the distribution network from the load forecasting module. This data includes the predicted load power for each time period within the control cycle. The duration and granularity of the control cycle are preset by the dispatch center. The duration of the control cycle can be one hour, four hours, or twenty-four hours, and the granularity of the time periods can be fifteen minutes or thirty minutes.

[0218] Step 312: Obtain the reference power supply of the distribution network in each time period within the control cycle.

[0219] The benchmark power supply characterizes the power that the distribution network can provide without the use of industrial adjustable resources. The benchmark power supply is obtained by superimposing the output of conventional power sources and the predicted output of new energy sources.

[0220] Step 313: Calculate the power gap between the predicted load power and the baseline power supply, and use the power gap as the target adjustment power for each time period.

[0221] The difference between the predicted load power and the baseline power supply for each time period is used as the power gap for that time period. A positive power gap indicates insufficient power supply, requiring industrial adjustable resources to reduce load; a negative power gap indicates excess power supply, requiring industrial adjustable resources to increase load absorption. The power gap is used as the target adjustment power for each time period, which is the future demand that needs to be tracked in the control objective function.

[0222] Step 320: Establish a control objective function that takes into account the system load tracking accuracy, operating costs, and risk costs, and set constraints.

[0223] Specifically, step 320 includes steps 321 to 325:

[0224] Step 321: Establish the system load tracking accuracy item.

[0225] Traditional control methods use only the absolute value of the difference between the control power and the target value as an indicator of tracking accuracy, without distinguishing the control reliability of different resources. This application introduces a resource priority index to weight the tracking accuracy term, allowing the control contribution of high-priority resources to account for a larger share in the tracking accuracy calculation.

[0226] The system load tracking accuracy term characterizes the degree of matching between the regulated power and the target regulated power. For each time period, the planned regulated power of each industrial adjustable resource is calculated as the product of the corresponding resource priority index, and then summed to obtain the weighted regulated power. The deviation between the weighted regulated power and the target regulated power is calculated, and the squared deviation is used as the single-time component of the system load tracking accuracy term. The single-time components of all time periods are accumulated to obtain the system load tracking accuracy term.

[0227] Step 322: Establish the runtime cost item.

[0228] The operating cost term measures the economic expenditure incurred by activating industrial adjustable resources to perform adjustments. The adjustment cost function is retrieved from the standardized contract templates of each industrial adjustable resource, and the planned adjustment power is substituted into the adjustment cost function to obtain the adjustment cost of a single industrial adjustable resource. The adjustment costs of all industrial adjustable resources over all time periods are summed to obtain the operating cost term.

[0229] It should be noted that the baseline cost and slope coefficient in the adjustment cost function are derived from the standardized contract template. The adjustment cost function parameters differ for different industrial adjustable resources, and the slope coefficient for metallurgical loads is usually higher than that for air conditioning loads.

[0230] Step 323: Establish risk cost items.

[0231] The potential range represents the upper and lower limits of the adjustable power of industrial resources. If the planned adjustment power exceeds the potential range, the actual implementation may fail to reach the planned value, causing the control effect to deviate from expectations. Traditional control methods treat the potential boundary as a hard constraint; once violated, the plan becomes infeasible. This application introduces a risk cost term as a soft constraint correction method, allowing the planned adjustment power to deviate from the potential range to a certain extent, but requiring the incurring corresponding risk costs.

[0232] The risk cost term is a correction term for deviations from the potential range. For each industrial adjustable resource in each time period, the deviation of the planned adjustment power from the potential range is calculated. If the planned adjustment power is between the lower and upper potential limits, the deviation is zero; if the planned adjustment power is less than the lower potential limit, the deviation is the difference between the lower potential limit and the planned adjustment power; if the planned adjustment power is greater than the upper potential limit, the deviation is the difference between the planned adjustment power and the upper potential limit. The deviation is multiplied by a preset risk penalty parameter and squared to obtain the risk cost component of a single industrial adjustable resource in a single time period. The risk cost components of all industrial adjustable resources in all time periods are summed to obtain the risk cost term.

[0233] Step 324: Establish the target function for regulation.

[0234] The system load tracking accuracy term, operational cost term, and risk cost term are weighted and summed to obtain the control objective function. The optimization objective of the control objective function is minimization. The weighting parameters of the system load tracking accuracy term, operational cost term, and risk cost term are preset by the dispatch center level based on the current power grid operating status and dispatch strategy. The weighting parameter of the system load tracking accuracy term is increased when emphasizing regulation effect; the weighting parameter of the operational cost term is increased when emphasizing economy; and the weighting parameter of the risk cost term is increased when emphasizing safety margin.

[0235] Step 325: Set constraints.

[0236] The potential range is used as a constraint condition for the control objective function, and response time constraints, node voltage constraints, and regional power balance constraints are set.

[0237] The potential range constraint requires that the planned adjustment power of each industrial adjustable resource in each time period does not exceed the lower and upper limits of the potential recorded in the potential certificate. Since the risk cost term has already penalized deviations from the potential range, the potential range constraint is treated as a soft constraint in this application, allowing for moderate overshooting during the optimization process.

[0238] The response time constraint requires that the time interval between the start time of the adjustment action of each industrial adjustable resource and the time of the issuance of the control command shall not be less than the allowable response delay time recorded in the standardized contract template, and the duration of the adjustment action shall not exceed the duration recorded in the standardized contract template.

[0239] The node voltage constraint requires that the voltage amplitude of each node after adjustment be between the allowable upper and lower voltage limits. The voltage sensitivity parameters of each node are obtained from the distribution network power flow calculation module. Based on the voltage sensitivity parameters and the planned adjustment power, the node voltage change after adjustment is calculated. The node voltage change is then superimposed onto the node voltage reference value before adjustment, and the superimposed node voltage is verified to meet the upper and lower voltage limit constraints.

[0240] The regional power balance constraint requires that the sum of the planned adjustment power of all industrial adjustable resources within each regional control unit equals the regional adjustment power command issued by the dispatch center to the regional control unit. The regional adjustment power command is allocated by the dispatch center based on the proportion of each regional control unit's average adjustable potential to the overall network's average adjustable potential.

[0241] Step 330: Solve the objective function to generate the optimal control scheme and write it into the blockchain smart contract.

[0242] Specifically, step 330 includes steps 331 to 334:

[0243] Step 331: Perform optimization on the control objective function to obtain the optimal planned control power of each industrial adjustable resource in each time period.

[0244] A quadratic programming method is used to optimize the control objective function. The system load tracking accuracy term and risk cost term in the control objective function are both quadratic functions of the planned regulation power, while the operating cost term is a linear function of the planned regulation power. The overall control objective function conforms to the standard form of a quadratic programming problem. The potential range constraints, response time constraints, node voltage constraints, and regional power balance constraints are transformed into linear inequality constraints and linear equality constraints, and then solved using a quadratic programming solver to obtain the optimal planned regulation power for each industrial adjustable resource in each time period.

[0245] Step 332: Generate the optimal control scheme based on the optimal plan to adjust the power.

[0246] The optimal planned adjustment power of each industrial adjustable resource in each time period is summarized to form the optimal control scheme. The optimal control scheme includes five fields: resource identifier, time period identifier, optimal planned adjustment power, adjustment direction, and execution time. The resource identifier is a unique number for the industrial adjustable resource, consistent with the resource identifier in the potential certificate and standardized contract template. The time period identifier is the sequence number of the time period within the control cycle. The optimal planned adjustment power is the power value obtained from the optimization solution. The adjustment direction identifier is reduction or increase; reduction indicates that the industrial adjustable resource needs to lower its load power, and increase indicates that the industrial adjustable resource needs to raise its load power. The execution time is the timestamp of when the industrial adjustable resource begins to execute the adjustment action.

[0247] Step 333: Calculate the hash digest of the optimal control scheme and write it into the blockchain smart contract.

[0248] The hash digest of the optimal control plan is calculated, and the hash digest, generation timestamp, and digital signature at the scheduling center level are written into the blockchain smart contract. The blockchain smart contract assigns a unique on-chain index to the optimal control plan, and this unique on-chain index is stored in association with the unique on-chain index of the standardized contract template for easy traceability during subsequent settlement.

[0249] Step 334: Issue the optimal control plan to each regional control unit and industrial adjustable resources through blockchain smart contracts.

[0250] The blockchain smart contract, based on the resource identifiers in the optimal control plan, breaks down the optimal control plan into regional control sub-plans for each regional control unit. After receiving the regional control sub-plan, the regional control unit further breaks it down into execution instructions for each industrial adjustable resource and sends them to the terminal resource level. Upon receiving the execution instructions, the industrial adjustable resources execute the adjustment actions according to the execution time and the optimal plan's power adjustment.

[0251] It is worth mentioning that the optimal control plan is issued through a blockchain smart contract, allowing all participants to verify the completeness and authenticity of the plan's content, thus avoiding the risk of instruction tampering present in traditional centralized scheduling platforms. Step 400: Collect the actual execution power of the response to the optimal control plan, execute settlement through the blockchain smart contract and calculate the execution fulfillment rate, and update the standardized contract template and resource priority index using the actual execution power and execution fulfillment rate.

[0252] In some embodiments, step 400 includes steps 410 to 430:

[0253] Step 410: Collect the actual execution power of the optimal control scheme.

[0254] It should be noted that the terminal data acquisition devices are deployed at the metering points of each industrial adjustable resource. These devices include smart meters and power transmitters. When the execution time recorded in the optimal control scheme arrives, the terminal data acquisition devices sample the instantaneous power of the industrial adjustable resource according to a preset sampling interval. An arithmetic average is performed on the real-time power data from all sampling points within each time period to obtain the actual execution power of the industrial adjustable resource within that time period. The terminal data acquisition devices upload the actual execution power to the regional control level through the node gateway. The regional control level then aggregates the data and reports it to the dispatch center level.

[0255] Step 420: Execute settlement via blockchain smart contract.

[0256] Specifically, step 420 includes steps 421 to 423:

[0257] Step 421: Call the standardized contract template through the blockchain smart contract.

[0258] The blockchain smart contract retrieves standardized contract templates for various industrial adjustable resources from the blockchain based on the resource identifiers in the optimal control scheme. The unique on-chain index of the standardized contract template and the unique on-chain index of the optimal control scheme have been associated and stored in step 330. The blockchain smart contract retrieves the standardized contract template through the association relationship of the unique on-chain indexes.

[0259] Step 422: Calculate the deviation between the actual power output and the planned power output indicated by the optimal control scheme.

[0260] For each industrial adjustable resource in each time period, the actual execution power is read from the execution record dataset, and the optimal planned adjustment power is read from the optimal control scheme. The difference between the actual execution power and the optimal planned adjustment power is obtained as the deviation value. A positive deviation value indicates that the actual adjustment exceeds the planned adjustment, while a negative deviation value indicates that the actual adjustment does not reach the planned adjustment.

[0261] Step 423: Calculate the settlement interaction value based on the adjustment cost function and deviation value in the standardized contract template.

[0262] The baseline cost and slope coefficient of the adjustment cost function are read from the standardized contract template. The optimal planned adjustment power is substituted into the adjustment cost function to obtain the planned settlement baseline value. The planned settlement baseline value is adjusted according to the deviation value to obtain the settlement interaction value. When the deviation value is zero, settlement is based on the planned settlement baseline value; when the deviation value is positive, compensation is given for the excess; when the deviation value is negative, the shortfall is deducted. The blockchain smart contract writes the settlement interaction value and the deviation value onto the chain, forming an immutable settlement certificate.

[0263] Furthermore, the execution performance rate is calculated based on the ratio of the deviation value to the planned power. Specifically, the execution performance rate measures the overall performance of industrial adjustable resources within the control cycle. For each industrial adjustable resource, the deviation value for all time periods within the control cycle is statistically analyzed, and the execution performance rate is calculated based on the ratio of the deviation value to the optimal planned adjustment power. The execution performance rate ranges from 0 to 1; a higher execution performance rate (closer to 1) indicates better performance.

[0264] Step 430: Update the standardized contract template and resource priority index using actual execution power and execution fulfillment rate.

[0265] Specifically, step 430 includes steps 431 to 433:

[0266] Step 431: Calculate the mean deviation and the deviation fluctuation based on the deviation value, and combine the mean deviation and the deviation fluctuation to form the execution error parameter.

[0267] For each industrial adjustable resource, the mean deviation and deviation fluctuation are calculated based on the deviation values ​​for each time period within the control cycle. The mean deviation reflects the direction and magnitude of the systematic execution deviation, while the deviation fluctuation reflects the degree of execution stability. The mean deviation and deviation fluctuation are combined to form the execution error parameter.

[0268] Step 432: Use the execution error parameter to correct the potential uncertainty parameter in the standardized contract template.

[0269] The deviation volatility is extracted from the execution error parameters, and the execution volatility is calculated based on the ratio of the deviation volatility to the potential mean parameter. Based on the preset correction smoothing parameters, a weighted average is performed on the original potential uncertainty parameter and the execution volatility to obtain the corrected potential uncertainty parameter, which is then written into the standardized contract template and the on-chain storage is updated.

[0270] The correction smoothing parameter is preset by the scheduling center level according to the control scenario, and the value of the correction smoothing parameter ranges from 0 to 1. The larger the correction smoothing parameter, the greater the proportion of execution volatility in the corrected potential uncertainty parameter, and the smaller the impact of the original potential uncertainty parameter; the smaller the correction smoothing parameter, the higher the degree of preservation of the original potential uncertainty parameter, and the smaller the impact of execution volatility on the correction result.

[0271] It should be noted that if the volatility of industrial adjustable resources remains high over multiple control cycles, the revised potential uncertainty parameter will gradually increase, leading to a greater penalty for industrial adjustable resources by the risk cost term, thereby reducing the likelihood that industrial adjustable resources will be selected in the optimal control scheme.

[0272] Step 433: Based on the preset memory decay parameter and the execution fulfillment rate, iteratively update the resource priority index.

[0273] A preset memory decay parameter controls the degree to which the historical resource priority index is retained during iterative updates. The value of the memory decay parameter ranges from 0 to 1. Based on the memory decay parameter, a weighted average is performed on the historical resource priority index and the execution fulfillment rate to obtain the updated resource priority index. The updated resource priority index is written to the potential database for use in the next adjustment cycle.

[0274] By iteratively updating the resource priority index based on the execution performance rate, the resource priority index of industrial adjustable resources with high execution performance rates will gradually increase, while the resource priority index of industrial adjustable resources with low execution performance rates will gradually decrease, forming a closed loop of resource selection based on survival of the fittest. Based on the above steps, this application also includes the following embodiments:

[0275] See Figure 4This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. The electronic device 40 includes: a processor 41, a memory 42, and a computer program; wherein,

[0276] The memory 42 is used to store the computer program, and the memory may also be flash memory. The computer program is, for example, an application program or functional module that implements the above method.

[0277] The processor 41 is configured to execute the computer program stored in the memory to implement the various steps performed by the device in the above method. For details, please refer to the relevant descriptions in the preceding method embodiments.

[0278] Alternatively, the memory 42 can be either standalone or integrated with the processor 41.

[0279] When the memory 42 is a device independent of the processor 41, the device may further include:

[0280] Bus 43 is used to connect the memory 42 and the processor 41.

[0281] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for interval potential modeling and optimal regulation and control of power distribution grid industrial adjustable resources, characterized in that, The method comprises the following steps: extracting a multi-dimensional resource feature set based on industrial load operation data, classifying the multi-dimensional resource feature set by using self-organizing mapping and calculating corresponding potential intervals to generate a potential certificate, wherein the potential interval is a power interval corresponding to each category, and the potential certificate is an encapsulated set corresponding to each industrial adjustable resource, which includes a potential interval, a potential mean parameter, a potential uncertainty parameter and a potential level index; constructing a node feature vector, partitioning the power distribution network, establishing a hierarchical control architecture, generating a standardized contract template based on the potential certificate and calculating a resource priority index; the standardized contract template based on the potential certificate comprises: extracting the potential interval, the potential mean parameter and the potential uncertainty parameter from the potential certificate; obtaining a preset regulation generation value function and a response time window parameter; generating the standardized contract template by using the potential interval, the potential mean parameter, the potential uncertainty parameter, the regulation generation value function and the response time window parameter; wherein the regulation generation value function is constructed based on a benchmark generation value and a slope coefficient; and the response time window parameter includes an allowed response delay time and a duration; the calculation of the resource priority index comprises: extracting the potential level index and the reliability index from the potential certificate, and extracting the response speed index and the economic index from the standardized contract template; performing standardization processing on the potential level index, the reliability index, the response speed index and the economic index; performing weighted summation on each index after standardization processing based on a preset proportion parameter vector to obtain an initial priority index; calculating a risk correction parameter based on the potential interval, and performing correction on the initial priority index by using the risk correction parameter to obtain the resource priority index; the risk correction parameter is used to reduce the priority of the resource whose fluctuation amplitude of the potential interval exceeds a preset range; based on the hierarchical control architecture and the resource priority index, constructing a control target function with the future demand of the predicted power distribution network as the target and the potential interval as the constraint condition, solving to generate an optimal control scheme and writing it into a blockchain smart contract; collecting actual execution power in response to the optimal control scheme, performing settlement through the blockchain smart contract and calculating an execution compliance rate, and updating the standardized contract template and the resource priority index by using the actual execution power and the execution compliance rate.

2. The method of claim 1, wherein the industrial load operation data includes power curve data and process constraint parameters; the multi-dimensional resource feature set includes interruptability index, transferability index, operation cost index, process constraint intensity index, fluctuation index and mode stability index; the interruptability index represents the ratio of allowed period reduction energy to total energy consumption; the transferability index represents the ratio of transferable energy to total energy consumption; the operation cost index represents the marginal regulation generation value of unit regulation power; the process constraint intensity index represents the ratio of locked period to total period; the fluctuation index represents the ratio of power curve standard deviation to mean value. The mode stability index represents an entropy value of a cluster distribution of a daily operation curve.

3. The method of claim 2, wherein, The multi-dimensional resource feature set is extracted based on the operation data of the industrial load, and includes: statistical analysis and parameter extraction are performed on the power curve data and the process constraint parameters to obtain overall features of the industrial load, the overall features including a power mean parameter, a power dispersion parameter, an energy efficiency index, and a response failure rate; the power curve data is divided into a plurality of process stage intervals according to the process constraint parameters, and stage features of each process stage interval are collected, the stage features including a power mean parameter, a power dispersion parameter, a process minimum start-stop time, a process stage label, and a process constraint parameter set in the stage; the collected overall features and stage features are integrated to obtain a basic feature set, and the multi-dimensional resource feature set is calculated based on the basic feature set.

4. The method of claim 1, wherein, the multi-dimensional resource feature set is classified and a corresponding potential interval is calculated using a self-organizing map, and includes: a self-organizing map two-dimensional grid network is constructed, a Euclidean distance between the multi-dimensional resource feature set and each node in the self-organizing map two-dimensional grid network is calculated, and a node with the smallest distance is determined as an optimal matching node; the multi-dimensional resource feature set is mapped to the optimal matching node to form a category set; for each category in the category set, power samples in an allowed adjustment period are extracted from the operation data, a quantile method is used to calculate upper and lower limits of potential of the category to obtain the potential interval; based on the operation data, an energy consumption proportion of a single industrial adjustable resource is calculated, and the potential interval is distributed to the single industrial adjustable resource according to the energy consumption proportion.

5. The method of claim 4, wherein, the node feature vector is constructed, the power distribution network is partitioned, and a hierarchical control architecture is established, and includes: node electrical parameters of the power distribution network are obtained, the node electrical parameters including a node equivalent impedance, a voltage response degree parameter, and a node load power parameter; an average adjustable potential and a potential uncertainty parameter of resources in a node are calculated based on the potential certificate; a comprehensive feature vector is constructed, and a node feature matrix is obtained by performing normalization processing on the comprehensive feature vector, the comprehensive feature vector including the node electrical parameters, the average adjustable potential, and the potential uncertainty parameter; two-dimensional clustering is performed on the node feature matrix, distances of nodes to cluster centers in an electrical dimension and in a potential dimension are respectively calculated, and a comprehensive distance is determined, nodes are divided into corresponding regional control units according to the comprehensive distance; the hierarchical control architecture is established based on the regional control units; the regional control units include a dispatching center level, a regional control level, and a terminal resource level.

6. The method of claim 1, wherein, the control objective function is constructed based on the resource priority index and the potential interval, and is targeted at meeting a predicted future demand of the power distribution network, and specifically, a control objective function that takes into account system load tracking accuracy, operation cost, and risk cost is established. The system load tracking accuracy represents a matching degree of the adjusted power and the future demand; The operation cost is calculated based on an adjustment cost value function in the standardized contract template; and the risk cost is a correction term deviating from the potential interval. The potential interval is taken as a constraint condition of the control target function, and a response time constraint, a node voltage constraint and a regional power balance constraint are set.

7. The method of claim 1, wherein, The settlement is performed by the blockchain smart contract, including: calling the standardized contract template by the blockchain smart contract; calculating a deviation value between the actual execution power and a planned power indicated by the optimal control scheme; calculating a settlement interaction value based on an adjustment cost value function in the standardized contract template and the deviation value; calculating the execution performance rate based on a ratio relationship between the deviation value and the planned power.

8. The method of claim 7, wherein, The actual execution power and the execution performance rate are used to update the standardized contract template and the resource priority index, including: calculating a deviation mean and a deviation fluctuation amount based on the deviation value, and combining the deviation mean and the deviation fluctuation amount to form an execution error parameter; performing correction on a potential uncertainty parameter in the standardized contract template by using the execution error parameter; performing iterative update on the resource priority index according to a preset memory decay parameter and the execution performance rate.

Images