Park energy management and control method and system based on building energy consumption monitoring

Abstract

The application discloses a park energy management and control method and system based on building energy consumption monitoring, relates to the process control technical field of energy management, and executes edge-end monitoring deployment, constructs a park simulation body through park twinning, determines a target energy structure according to energy management emphasis, executes energy consumption overrun analysis and energy scheduling balanced decision-making, determines an energy management and control strategy and reversely transmits the energy management and control strategy to intelligent monitoring equipment for local management, and is used for solving the technical problems that the prior art cannot make light and flexible analysis and decision-making for diversified park energy scenes, leading to redundant processing modes and insufficient scene adaptability of energy management. The application realizes light and efficient analysis based on energy scenes and precise management of park energy management.

Description

Technical Field

[0001] This invention relates to the field of process control technology in energy management, specifically to a method and system for energy management and control of industrial parks based on building energy consumption monitoring. Background Technology

[0002] Current energy management systems in industrial parks generally employ a centralized architecture, which suffers from problems such as large response delays and low control precision. Although some improvement solutions attempt to introduce distributed computing or artificial intelligence algorithms, they still cannot simultaneously meet the multiple requirements of real-time performance, adaptability, and stability, thus hindering the improvement of energy management efficiency in industrial parks. They struggle to adapt to the dynamic changes in building loads, especially showing poor optimization results for nonlinear loads such as air conditioning and elevators. Existing technologies lack effective decoupling methods when dealing with multi-energy coupling, leading to high computational complexity and optimization results that deviate from reality.

[0003] In summary, existing technologies still suffer from the inability to perform lightweight and flexible analysis and decision-making for diverse energy scenarios in industrial parks, resulting in redundant processing methods and insufficient adaptability of energy management to various scenarios. Summary of the Invention

[0004] This application provides a method and system for park energy management based on building energy consumption monitoring, which is used to address the technical problems in existing technologies that cannot perform lightweight and flexible analysis and decision-making for diverse park energy scenarios, resulting in redundant processing methods and insufficient scenario adaptability of energy management.

[0005] In view of the above problems, this application provides a method and system for park energy management based on building energy consumption monitoring.

[0006] Firstly, this application provides a method for park energy management based on building energy consumption monitoring. The method includes: implementing edge-end monitoring deployment, wherein intelligent monitoring devices with remote transmission capabilities are deployed in a distributed manner to establish local recursive connections based on the park's local area and connect to the energy management platform; constructing a park simulation by creating a park twin, and performing a first decoupling based on monitoring guidance and a second iteration based on linear approximation on the park simulation according to the energy management focus, to determine the target energy structure, wherein the monitoring guidance includes at least item-zone-household, and the park simulation is updated with the network data received by the platform; for the target energy structure, performing energy consumption exceedance analysis and energy dispatch balancing decision-making, determining the energy management strategy and transmitting it in reverse to the intelligent monitoring devices for local management, wherein the energy slack of flexible loads is used as the decision target, and stable operation constraint management is carried out by introducing power angle stabilization control and tripping actions.

[0007] Secondly, this application provides a park energy management system based on building energy consumption monitoring. The system includes: a deployment module for performing edge-to-end monitoring deployment, wherein intelligent monitoring devices with remote transmission capabilities are deployed in a distributed manner to form a local recursive connection based on the park's local area and to connect to the energy management platform; a construction module for constructing a park simulation by performing park twins, and determining the target energy structure by performing a first decoupling based on monitoring guidance and a second iteration based on linear approximation according to the energy management focus, wherein the monitoring guidance includes at least item-zone-household, and the park simulation is updated with the network data received by the platform; and a decision module for performing energy consumption limit analysis and energy dispatch balance decision for the target energy structure, determining the energy management strategy and transmitting it back to the intelligent monitoring devices for local management, wherein the energy slack of flexible loads is used as the decision target, and stable operation constraint management is carried out by introducing power angle stabilization control and tripping actions.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The park energy management method based on building energy consumption monitoring provided in this application implements edge-end monitoring deployment. This involves distributed deployment of intelligent monitoring devices with remote transmission capabilities, establishing local recursive connections within the park's local area, and connecting to the energy management platform. A park simulation is constructed using a park twin. Based on the energy management focus, the simulation undergoes a first decoupling based on monitoring guidance and a second iteration based on linear approximation to determine the target energy structure. The monitoring guidance includes at least item-zone-household aspects. The park simulation updates with the network data received by the platform. For the target energy structure, energy consumption exceedance analysis and energy dispatch balancing decisions are performed to determine the energy management strategy, which is then transmitted back to the intelligent monitoring devices for local management. The energy slack of flexible loads is used as the decision target, and stable operation constraint management is implemented by introducing power angle stabilization and tripping actions. This method addresses the technical problems in existing technologies where lightweight and flexible analysis and decision-making for diverse park energy scenarios are impossible, leading to redundant processing methods and insufficient scenario adaptability of energy management. It achieves lightweight and efficient analysis based on energy scenarios and precise management of park energy. Attached Figure Description

[0010] Figure 1 This application provides a schematic diagram of the process for a park energy management method based on building energy consumption monitoring;

[0011] Figure 2 This application provides a flowchart illustrating the second iteration of the linear approximation-based energy management method for industrial parks based on building energy consumption monitoring.

[0012] Figure 3 This application provides a schematic diagram of the structure of a park energy management system based on building energy consumption monitoring.

[0013] Figure labeling: Deployment module 11, Construction module 12, Decision module 13. Detailed Implementation

[0014] This application provides a method and system for park energy management based on building energy consumption monitoring, which solves the technical problems in the existing technology that are unable to perform lightweight and flexible analysis and decision-making for diverse park energy scenarios, resulting in redundant processing methods and insufficient adaptability of energy management scenarios.

[0015] Example 1: As Figure 1 As shown, this application provides a method for park energy management based on building energy consumption monitoring, the method comprising:

[0016] S1: Execute edge-end monitoring deployment, which involves distributing intelligent monitoring devices with remote transmission capabilities to establish local recursive connections based on the park's local area and connect to the energy management platform for network access.

[0017] In this embodiment, intelligent monitoring devices with remote transmission capabilities are configured in each energy zone within the park using a distributed deployment method. For example, the park is divided based on building clusters, domains, etc., and intelligent monitoring devices are deployed for each domain. Preferably, the intelligent monitoring devices possess data acquisition, edge computing, and remote communication interaction capabilities.

[0018] Simultaneously, based on the park's local area network architecture, each intelligent monitoring device achieves data connectivity from the physical layer to the platform layer via wired or wireless communication protocols, such as LoRa, ZigBee, or Industrial Ethernet. This involves establishing local recursive connections, forming a hierarchical or mesh-topology data interaction network between devices. For example, floor-level devices upload data tier by tier to a regional aggregation node, with the deployed intelligent monitoring devices serving as the final aggregation point. The data is then further transmitted to the energy management platform via communication, achieving initial aggregation and processing of monitoring data at the edge. Specifically, the monitoring data carries spatiotemporal codes during its flow, including time and spatial encoding, for data traceability management. For example, B1F2-E01-0512 represents the reading of device E01 in Zone 2, Floor 1 of Building B at the specified timestamp 0512.

[0019] In a preferred embodiment of this application, the local recursive connection can use a heartbeat mechanism to maintain network stability, such as sending a liveness signal every 30 seconds. At the same time, the connection to the network implements an encrypted handshake to ensure the integrity and confidentiality of the monitoring data during transmission. The resulting edge-end collaborative monitoring system can simplify the interactive network, avoid redundant interactive threads, and ensure the orderliness and traceability of monitoring and transmission.

[0020] S2: By constructing a park simulation through park twins, and based on the energy management focus, the park simulation is subjected to a first decoupling based on monitoring guidance and a second iteration based on linear approximation to determine the target energy structure. The monitoring guidance includes at least item-zone-household. The park simulation is updated with the network access data received by the platform.

[0021] In this embodiment, the park simulation uses 3D modeling technology to map the physical park into a virtual entity containing building structures, energy equipment, and energy-consuming units; that is, the physical park and the park simulation are mirror images of each other. For example, a visualized energy system is constructed by overlaying BIM model and SCADA data. The park simulation is built and stored on the energy management platform. When the intelligent monitoring equipment uploads new energy consumption data, such as a surge in air conditioning unit power consumption during a certain period, the parameters of the park simulation are automatically adjusted to reflect the actual operating status.

[0022] Furthermore, as the park simulation is updated, it reflects the latest energy status of the park. This further identifies the key areas requiring energy management, improving the targeted nature of subsequent processing. For example, focusing on different types of energy structures, i.e., conducting structural analysis on the sub-items in the monitoring guidelines.

[0023] The monitoring guidance is specifically reflected in three dimensions: itemized, namely, divided by energy type such as electricity / gas / heat; zoned, namely, divided by building function such as office area / activity area; and unit-based, namely, divided by independent energy-consuming unit such as tenant meter.

[0024] In practical implementation, for example, when decoupling a commercial complex, the integrated energy network is first broken down into independent functional modules such as lighting systems, elevator groups, and refrigeration plants, forming several independent energy subsystems. This decoupling method effectively reduces the analytical complexity caused by multi-energy coupling, allowing each subsystem to be analyzed for energy consumption independently.

[0025] Furthermore, for the decoupled energy subsystem, the presence of some nonlinear energy sources makes it difficult to perform regular analysis and decision-making, thus increasing the analytical complexity. This application employs a second-iteration optimization based on linear approximation to transform it into a linear surrogate analysis.

[0026] Specifically, the first step is to identify nonlinear energy loads in the subsystem, such as variable frequency drive centrifugal chillers, and then perform linear approximation processing on them. For example, the COP curve is discretized into piecewise linear functions, transforming the nonlinear characteristics into computable linear surrogate characteristics. Preferably, after multiple iterative corrections, such as continuously adjusting the approximation coefficients using the least squares method, a target energy structure that reflects actual operating characteristics and is easy to optimize is finally obtained, serving as a benchmark model for subsequent management and control strategy formulation.

[0027] Furthermore, after implementing the edge-end monitoring deployment, step S2 of this application includes:

[0028] The intelligent monitoring device integrates monitoring data from locally recursively connected monitoring devices to determine an energy dataset, wherein each monitoring data is identified by a device spatiotemporal code; the intelligent monitoring device then connects the energy dataset to the network and transmits it back to the energy management platform to update the park simulation.

[0029] In this embodiment, during the operation of the edge-to-end monitoring architecture, the intelligent monitoring device, acting as a data aggregation node, first integrates and processes the raw data collected by locally recursively connected lower-level monitoring devices, such as ammeters installed in end-point distribution boxes. Specifically, each monitoring data packet is assigned a unique device spatiotemporal code, i.e., an identifier that combines location and time coding, ensuring the traceability of data in the spatiotemporal dimension. In a feasible embodiment, this coding system is automatically generated through the device's built-in positioning module and clock chip.

[0030] Furthermore, in the data integration phase, intelligent monitoring equipment integrates multi-source data and generates standardized energy datasets according to a preset data structure, including fields such as load values, quality flags, and time synchronization markers. For example, a smart meter integrates parameters such as active power, reactive power, and voltage harmonics collected at 15-minute intervals into an MMS message conforming to the IEC 61850 standard.

[0031] Preferably, during this process, data validity verification is performed simultaneously. For example, transmission errors are detected based on CRC checksums, and abnormal data with a mutation rate exceeding 20% ​​are identified and filtered for outliers through a sliding window to ensure the validity of the data uploaded to the platform.

[0032] Furthermore, the intelligent monitoring equipment uploads energy datasets to the energy management platform through the established network access channel. Upon receiving the data, the platform immediately triggers an update mechanism for the park simulation: specifically, it first parses the device's spatiotemporal code to determine the virtual device node to which the data belongs, for example, mapping A2-5F-Z3-AC to the corresponding air conditioning object in the park simulation. Then, it injects real-time monitoring values, such as the current cooling power of 32.5kW, into the corresponding parameter library. Simultaneously, the platform records the data version number, using a UTC timestamp as the version identifier, to support historical state retrospection.

[0033] Furthermore, step S2 of this application includes: the energy management focus is at least one of item-zone-household; and the park simulation is decoupled according to the energy management focus.

[0034] If energy management focuses on individual items, the decoupling method is as follows: determine the comprehensive energy network of the park simulation body, decompose the comprehensive energy network according to multiple independent energy functions, and determine multiple functionally independent energy networks as the first decoupling result.

[0035] In this embodiment, the focus of energy management is determined by selecting at least one dimension from item-based, zone-based, or household-based management as the core decoupling basis according to actual control needs. For example, item-based management is preferred in industrial park scenarios, while a zone-household combination model may be adopted in commercial complex scenarios.

[0036] The energy management focus is set automatically through the platform configuration interface and embedded as metadata in the topology description file of the park simulation. When the focus is set to a sub-dimensional aspect, a specific decoupling process is initiated: First, the integrated energy network established in the park simulation is traversed, that is, the interwoven network structure with different energy structures and different energy locations. This is done by identifying energy conversion nodes in the network, such as electric chiller units as the conversion interface between electrical energy and cold energy, and energy transmission paths, such as the power supply circuit of the low-voltage distribution bus and the air conditioning system.

[0037] Furthermore, when performing decoupling operations, independent energy functions can be divided according to the physical characteristics of energy flow. For example, electricity distribution-lighting electricity-lighting control can be identified as a complete functional chain, and the integrated energy network can be decomposed into several functional sub-graphs.

[0038] Preferably, in the embodiments of this application, decoupling is performed based on the functional chain.

[0039] Specifically, the integrated energy network is traversed, and for each energy node in the network architecture, functional boundary positioning is performed to determine the complete functional chain of the energy node's operation process. This involves determining the functional chain based on the energy allocation rules, usage mechanisms, underlying control logic, energy dispatch paths, and energy flow attributes of each energy node, thus integrating the complex integrated energy network into functional chains based on each energy node.

[0040] Furthermore, the integrated energy network is first decomposed according to the type of energy node, such as lighting control type, equipment control type, etc. Then, the decomposed sub-energy network is reconstructed using a functional chain representing an energy node as a reconstruction node to determine several functional subgraphs, wherein the number of functional subgraphs is consistent with the number of energy node types.

[0041] In practice, the functional boundary points in the network are first marked, such as the transformer secondary side outgoing switch as the power distribution functional boundary and the chilled water supply and return water pipe electric valve as the cooling functional boundary. Then, virtual breakpoint operations are performed at the boundary points to ultimately generate multiple uncoupled energy subsystems, such as independent power supply subsystems, cooling subsystems, lighting subsystems, etc.

[0042] Simultaneously, each subsystem retains complete energy flow calculation attributes, such as load characteristics, forming the first decoupling result from these functionally independent energy networks. Preferably, the mapping relationship established during the decoupling process—that is, the mapping used to record the affiliation between the integrated energy network and the decoupled subsystem networks—is simultaneously updated and temporarily stored, providing a topological basis for subsequent reverse reconstruction.

[0043] Furthermore, such as Figure 2 As shown, a second iteration based on linear approximation is performed. Step S2 of this application includes:

[0044] Based on the first decoupling result, linear energy loads and nonlinear energy loads are divided; for each nonlinear energy load, an uncertain linear probability is introduced to perform linear approximation processing to determine a proxy linear energy load; based on the proxy linear energy load, the nonlinear energy loads in the first decoupling result are replaced to determine the target energy structure.

[0045] In this embodiment, after obtaining the first decoupling result, load characteristic analysis is first performed on each functionally independent energy network. For example, by loading a historical operating database of the equipment, such as storing load curve data for at least one full year, feature extraction is used to determine the load pattern, classifying linear energy loads into linear and nonlinear energy loads. Linear energy loads refer to equipment whose power-time curves follow a linear relationship, such as LED lighting systems with constant illuminance; while nonlinear energy loads include equipment with various time-varying characteristics, such as variable frequency drive centrifugal chillers, whose power characteristics exhibit segmented, periodic, or random fluctuations.

[0046] In one feasible embodiment, by setting a judgment threshold, such as determining nonlinearity if the linear correlation coefficient is less than 0.85, the classification and labeling are automatically completed and displayed in different colors, such as blue representing linear energy load and red representing nonlinear energy load.

[0047] Furthermore, for the identified nonlinear energy loads, a probabilistic linear approximation process is implemented. In one feasible embodiment, an uncertain linear probability is first established, comprising three core parameters: a linear approximation interval, a confidence level, and an approximation error band. Taking a variable frequency air conditioner as an example, its actual COP curve is discretized into several operating points, such as the energy efficiency points corresponding to 20%, 50%, and 100% load rates. Within each operating point segment, an approximation process is performed using a linear approximation interval, a confidence level greater than or equal to 0.95, and an approximation error band less than or equal to 5% as criteria to determine the proxy linear energy load. This process retains the main dynamics of the original nonlinear characteristics, such as the response trend during load abrupt changes, while also possessing the computability of a linear system.

[0048] Preferably, the nonlinear curve based on the time series is first discretized based on the working point, and then the adjacent working points are determined to be in intervals, such as 20% to 25% of the working points. When performing linear approximation later, the linear approximation process is performed on the multiple intervals to ensure the linear processing effect.

[0049] Finally, after completing the linear proxy analysis of all nonlinear loads, the nonlinear energy loads in the first decoupling result are replaced. Based on the proxy linear energy loads and the linear energy loads, the energy network is reconstructed to determine the target energy structure. This preserves the main dynamics of the original nonlinear characteristics while maintaining the computability of a linear system, laying a solid foundation for subsequent energy management analysis.

[0050] S3: For the target energy structure, perform energy consumption limit analysis and energy dispatch balance decision, determine energy management strategy and transmit it in reverse to the intelligent monitoring equipment for local management. Among them, the energy relaxation of flexible load is used as the decision target, and stable operation constraint management is carried out by introducing power angle stabilization control switching action.

[0051] In this embodiment, after obtaining the target energy structure, a dynamic limit-crossing determination mechanism is first established to determine whether the energy consumption of each energy node exceeds the limit. In one specific embodiment, based on real-time updated monitoring data, including load data of each node at a 15-minute granularity, a preset energy consumption standard library is loaded, which stores time-sharing and zone-specific energy consumption thresholds, such as a daytime power limit of 85kW for the office area on weekdays. A two-layer determination is used to perform limit-crossing analysis. Specifically, the first layer performs static threshold comparison, that is, directly compares the measured value with the standard value. The second layer performs dynamic threshold analysis, that is, compares the energy consumption fluctuation trend. When any layer triggers the limit-crossing condition, the energy node is marked as being in an limit-crossing state, as shown by a flashing red alarm icon in the topology diagram.

[0052] Further energy dispatching and balancing decisions are made for the over-limit parts. Rigid loads need to meet their adjustment standards, and flexible loads need to meet their energy relaxation. The dual optimization objectives are to minimize total energy consumption and balance regional loads. Based on these constraints, energy decision analysis based on low energy consumption is carried out. Under the constraint of satisfying the power supply priority of rigid loads, such as elevators and security systems, which must be 100% guaranteed, the optimal dispatching scheme is solved as the energy management strategy.

[0053] Furthermore, the energy management strategy is decoupled to correspond to different load sides, and the management strategy for different loads is determined. This strategy is then transmitted from the energy management platform to the intelligent monitoring device that is recursively connected to the load for local response management.

[0054] Meanwhile, the process introduces a power angle stabilization strategy as a dynamic constraint: by monitoring the power angle difference at the grid connection point in real time, for example, when an instability risk of greater than or equal to 5° is detected, a preset power cut-off sequence is automatically activated, such as cutting off 30%, 50%, and 100% of non-critical loads in stages.

[0055] Furthermore, before performing energy consumption exceedance analysis and energy dispatch balance decision-making, step S3 of this application includes:

[0056] For real-time campus energy scenarios, flexible loads and rigid loads are identified, with each flexible load marked with an energy slack; based on the flexible loads and the rigid loads, the target energy structure is marked.

[0057] In this embodiment, during the real-time energy scenario processing stage, the load type is first identified through equipment characteristics and real-time operating data. Specifically, the process involves retrieving inherent attribute records from the equipment registration database, such as equipment technical parameters provided by the manufacturer and the intended use classification indicated during installation. This is combined with analysis of the current operating mode, such as identifying whether the elevator is in normal operation or energy-saving mode based on work cycle characteristics, and determining the rigidity and flexibility of the load.

[0058] Rigid loads are defined as critical equipment that must be continuously powered, such as fire protection systems and data center UPS power supplies. Their criteria include whether they involve personal safety, whether any form of power interruption is allowed, and whether they have strict quality requirements, such as voltage fluctuation rate ≤2%. Flexible loads, on the other hand, are labeled as adjustable energy consumption units, such as landscape lighting and non-production air conditioning. Their typical characteristics include adjustable operating parameters and the ability to allow short-term interruptions or derating.

[0059] For the categorized flexible loads, their energy slack index is further calculated, which represents the adjustable degrees of freedom, such as current operating requirements and external environmental parameters. The energy slack range is 0-1, measuring the freely adjustable energy slack of each flexible load. For example, in a variable air volume (VAV) air conditioning system in an office area, if the current load rate is 65% during the transition season and the equipment has the potential to reduce the load by 40%, then the energy slack is 0.4. Preferably, a sliding time window mechanism is used, for example, updating every 5 minutes, to ensure that the slack index dynamically reflects the latest adjustable capabilities of the equipment.

[0060] Finally, after completing the load classification and relaxation labeling, different identifiers were used to mark the flexible and rigid loads in the target energy structure, and the energy relaxation of the flexible loads was associated, laying a solid foundation for subsequent energy management analysis.

[0061] Furthermore, in performing energy consumption exceedance analysis and energy dispatch balance decision-making, step S3 of this application includes:

[0062] For the energy consumption standards of the park, set static and dynamic energy consumption thresholds on the load side; use the static and dynamic energy consumption thresholds to determine the target energy structure, identify the energy structure that exceeds the limit, and make energy scheduling and balancing decisions for the energy structure that exceeds the limit.

[0063] In this embodiment, during the energy consumption limit judgment stage, preset static and dynamic thresholds are loaded from an energy consumption standard database to determine energy consumption. The static thresholds are derived from equipment rated parameters and energy consumption standards, such as the power limit per unit area specified in the "Energy Conservation Design Standard for Public Buildings" GB50189, and are in fixed numerical form, such as the lighting power density limit for office areas being 9W / m². 2 The dynamic threshold is generated based on historical energy consumption pattern analysis, using dynamic benchmark values ​​under time series, such as continuous load increments on the load side, and the target energy structure is determined based on the two types of thresholds.

[0064] In a feasible embodiment, when performing limit violation determination, a hierarchical scanning method is used to traverse and analyze the target energy structure: In the first-level scan, the real-time energy consumption data of each node is directly compared with the static threshold. For example, if the current value of a certain distribution circuit exceeds the rated capacity of the circuit breaker, nodes that obviously exceed the limit are immediately marked. In the second-level scan, the deviation between the current energy consumption trend and the dynamic threshold is calculated in combination with the time-series context information, and whether the current load change rate will touch the upper limit of the dynamic threshold. Preferably, the degree of limit violation is measured in levels, for example, divided into three levels: mild, moderate, and severe, corresponding to exceedance ranges of 5%, 10%, and 15%, respectively. For example, when the total power of the chiller station is detected to exceed both the static threshold (equipment safety limit) and the dynamic threshold (current time period prediction value), the energy subsystem is marked as severely exceeding the limit, and the relevant power supply path is highlighted in the target energy structure to identify the energy structure that exceeds the limit.

[0065] Furthermore, for energy dispatching decisions regarding over-limit energy structures, firstly, based on previously marked load characteristic information, the mandatory fulfillment of rigid loads and the regulation of dispatchable flexible parts are identified from the over-limit areas. Decisions are made with low power consumption and energy balance as the guiding principles to determine the energy management strategy.

[0066] Furthermore, for the aforementioned over-limit energy structure, energy dispatch equilibrium decision-making is performed. Step S3 of this application includes:

[0067] For the aforementioned over-limit energy structure, rigid loads are identified as the first priority, and flexible loads are identified and an optimization space is built using energy relaxation. Based on the first priority and the optimization space, with low energy consumption as the first objective and energy balance as the second objective, multi-objective optimization iterative decision-making is carried out to determine the energy management strategy, wherein the park microgrid is integrated into the distribution network for coordinated energy supply.

[0068] In this embodiment, during the optimization decision-making stage of the energy structure exceeding the limits, the first step is to identify critical loads and manage resources: by traversing and scanning the energy network nodes in the exceeding-limit area, based on the preset load criticality level labels (i.e., priority levels 1-5 marked during equipment registration, where level 1 is life safety related equipment), all rigid loads are extracted to form a set of loads that must be guaranteed. For example, fire pumps are marked as level 1, and data center cooling systems are marked as level 2. These rigid loads, as unadjustable boundary conditions, must have their power supply requirements forcibly met.

[0069] Simultaneously, all flexible loads within the region are retrieved, and a multi-dimensional optimization space is constructed based on their real-time updated energy relaxation. Geometrically, each flexible load corresponds to an adjustable interval in the solution space. For example, if the energy relaxation of an air conditioning unit is 0.3, then its adjustable power range is the adjustable region of the solution space. The adjustable intervals of all flexible loads constitute the feasible region of the optimization problem.

[0070] Further, based on the aforementioned constraints and optimization space, energy management decisions are made with at least two optimization objectives: minimizing total energy consumption and emphasizing energy balance. In a feasible embodiment, a basic energy-saving method based on minimum energy consumption is quickly solved using linear programming to find the solution with the smallest load rate difference among distribution branches under the energy relaxation constraint. The basic energy-saving method based on minimum energy consumption and the solution with the smallest load rate difference among distribution branches are the optimal solutions with a single focus. The two focuses, minimizing energy consumption and balancing energy, are then considered simultaneously. Specifically, the two are mapped; if the mapping is consistent, it indicates that the solution meets both objectives, meaning it is optimal in both focuses. For solutions that do not meet the objectives, iterative optimization is used for further optimization. Based on the conventional optimization method, a two-step process is performed: single-objective focus decision-making and consistency checking filter out solutions with optimal focuses. For inconsistent solutions, further optimization is performed by balancing total energy consumption and energy balance. This approach effectively reduces the amount of data required for optimization, ensuring the quality of the optimized solution with minimal computing power and a relatively small number of iterations. Specifically, a preset adjustment step size is determined, and each energy node is randomly perturbed and adjusted. Its energy consumption and energy balance are measured, and this process is iterated until the preset number of iterations is met. The solution with the lowest energy consumption and highest energy balance among the iterative solutions is then selected as the optimized solution. The energy management strategy is then derived by combining the optimized solution with the mapped portion of the solution.

[0071] In this embodiment, the park's energy supply is achieved through a combination of the park's microgrid and distribution network. Special consideration is given to the interactive effects of microgrid grid-connected operation: real-time monitoring of the power exchange value at the point of common coupling (PCC) is conducted, and voltage-power angle stability is introduced to ensure that the adjustment scheme does not cause grid oscillations. For example, when a sudden change in the output of a photovoltaic inverter is detected, causing the power angle difference at the PCC point to approach a critical value, the damping coefficient constraint of that node is automatically increased.

[0072] Furthermore, by introducing power angle stabilization control for machine cutting action to achieve stable operation constraint management, step S3 of this application includes:

[0073] A stability control database is introduced, which contains multiple sequences representing fault scenarios and tripping actions; based on the stability control database, energy operation risk control analysis is performed to generate stability control tripping instructions; the stability control tripping instructions are incorporated into the energy management strategy.

[0074] In this embodiment, during the stability control strategy integration stage, a pre-built stability control database is first loaded. This database adopts a fault scenario-switching mechanism architecture and mines the mapping relationship between stored historical fault cases and response strategies. Specifically, each record entry contains three structured data groups: a fault feature vector, such as a feature space composed of parameters like voltage drop amplitude [0.1, 0.9] and frequency change rate [0.1, 5]; a device state matrix, such as a snapshot of the operating conditions of each key device at the time of the fault; and a verified switching action sequence, such as a set of device operation instructions sorted by action priority. Through scenario matching, corresponding mapping relationships are established and aggregated into the stability control database.

[0075] Furthermore, real-time monitoring of stability indicators, including but not limited to: bus voltage deviation rate, power angle difference Δθ, frequency fluctuation rate, etc., is performed using a sliding time window, typically set to 200ms, to calculate the gradient changes of each indicator. When any indicator exceeds a preset warning threshold, the fault scenario under that indicator state is determined, and scenario matching is performed in the stability control database. If a match is successful, the corresponding tripping action template is invoked, such as immediately tripping 30% of non-critical loads → if recovery is not achieved after 200ms, an additional 20% tripping is implemented in a tiered strategy. Simultaneously, verification is performed using the park simulation, by rehearsing the strategy execution effect in a virtual environment and calculating key indicators such as expected recovery time. This determines the stability control tripping command. The stability control tripping command is further incorporated into the energy management strategy for park energy management.

[0076] Furthermore, determining the energy management strategy and transmitting it in reverse to the intelligent monitoring equipment for local management, step S3 of this application includes:

[0077] Based on the deployment and local recursive connection relationship of the intelligent monitoring device, the energy management strategy is decomposed to determine N sub-strategies, wherein the N sub-strategies correspond to the intelligent monitoring device; the N sub-strategies are then transmitted in reverse to the intelligent monitoring device for distributed energy management guidance.

[0078] In this embodiment, during the strategy decomposition and distribution phase, the topological scope of the energy management strategy is first analyzed, i.e., the energy nodes used to respond to different parts of the strategy. The specific execution process is as follows: based on the device influence range parameters in the strategy instructions, such as the B2-5F air conditioning group, the communication path of the controlled devices is traced back in the park's digital twin model, such as: energy management platform → intelligent monitoring equipment → floor switch → terminal air conditioning controller, etc.

[0079] The energy management strategy, i.e., the global control strategy, is further divided into N sub-strategies: for continuous adjustment commands, such as temperature setpoint adjustments, spatial division is performed according to the equipment's jurisdiction, such as allocating the overall energy-saving target to each area controller according to area ratio; for discrete control commands, such as equipment start-up and shutdown, they are directly mapped to specific actuator nodes. During the decomposition process, the temporal dependencies between devices are carefully considered; for example, the chilled water pump can only run 5 minutes after the cooling tower starts. A task sequence is constructed using a directed acyclic graph to ensure the integrity of the execution logic of the sub-strategies.

[0080] Each sub-policy is encapsulated as a standardized information object, which, for example, includes: target device ID (identified by both IPv6 address and MAC address), execution time window (with valid start time and timeout parameters), expected effect value (e.g., power reduction of 15% ± 3%), and rollback conditions (e.g., terminating the policy when the ambient temperature exceeds 28°C).

[0081] The reverse transmission of sub-policies adopts a hierarchical confirmation mechanism: that is, the transmission is based on the architecture of energy management platform → intelligent monitoring equipment → floor switch → end controller, with the intelligent monitoring equipment gathered by the end controller as the transmission target, receiving the sub-policies and distributing them to lower-level controllers for execution.

[0082] The energy management method for industrial parks based on building energy consumption monitoring provided in this application has the following technical effects:

[0083] 1. Distributed coverage of data collection is achieved through intelligent monitoring devices with remote transmission capabilities, supporting local recursive networking. Spatiotemporal coding data identification is introduced. Local recursive connections reduce reliance on cloud transmission, improve real-time performance, and achieve a measured latency reduction of over 30%. Spatiotemporal codes support rapid location of energy consumption anomalies, providing a foundation for zoned and household management.

[0084] 2. The park's energy network is broken down into functionally independent sub-networks by item, zone, and household. A probabilistic linear proxy is introduced for nonlinear loads, simplifying computational complexity through iterative approximation. This enables lightweight analysis; linear approximation makes nonlinear system optimization possible, supporting high-precision scheduling. Energy relaxation is defined for flexible loads, constructing an optimization space. The optimization objectives are low energy consumption and balanced energy supply, solved in conjunction with rigid load priority constraints. Peak loads are reduced through flexible load adjustment; local overloads are avoided through supply-demand balance.

[0085] 3. Adopt the power system power angle stability control strategy and pre-set the generator tripping command sequence corresponding to the fault scenario. Trigger the stability control command during over-limit analysis to forcibly ensure system stability. The fault response time is reduced from seconds to milliseconds, avoiding cascading failures.

[0086] In summary, energy efficiency optimization is achieved through decoupled flexible scheduling using digital twins, while power system stability control technology is introduced to ensure a safety baseline. The edge-end architecture and linearization method are adaptable to campuses of different sizes; adding new equipment only requires expanding the recursive network. Power angle control of the power system is migrated to building energy management, solving dynamic stability problems that are difficult to handle with traditional methods. This solution is suitable for high-density load campuses, achieving comprehensive energy efficiency improvements with low power consumption while ensuring power supply reliability.

[0087] Example 2: Based on the same inventive concept as the park energy management method based on building energy consumption monitoring in the previous examples, such as... Figure 3 As shown, this application provides a campus energy management system based on building energy consumption monitoring, the system comprising:

[0088] Deployment module 11 is used to perform edge-end monitoring deployment, wherein intelligent monitoring devices with remote transmission capabilities are deployed in a distributed manner to make local recursive connections based on the park's local area and to connect to the energy management platform.

[0089] Module 12 is used to construct a park simulation by performing park twins, and to perform a first decoupling based on monitoring guidance and a second iteration based on linear approximation on the park simulation according to the focus of energy management, to determine the target energy structure, wherein the monitoring guidance includes at least item-zone-household, and the park simulation is updated with the network access data received by the platform.

[0090] Decision module 13 is used to perform energy consumption limit analysis and energy dispatch balance decision for the target energy structure, determine energy management strategy and transmit it in reverse to the intelligent monitoring equipment for local management. Among them, the energy relaxation of flexible load is used as the decision target, and stable operation constraint management is carried out by introducing power angle stabilization control switching action.

[0091] Furthermore, the construction module 12 is used to perform the following steps: the intelligent monitoring device integrates the monitoring data of the locally recursively connected monitoring devices to determine the energy dataset, wherein each monitoring data is identified by a device spatiotemporal code; according to the intelligent monitoring device, the energy dataset is connected to the network and transmitted back to the energy management platform to update the park simulation.

[0092] Furthermore, the construction module 12 is used to perform the following steps: the energy management focus is at least one of item-zone-household; and the park simulation is decoupled according to the energy management focus.

[0093] If energy management focuses on individual items, the decoupling method is as follows: determine the comprehensive energy network of the park simulation body, decompose the comprehensive energy network according to multiple independent energy functions, and determine multiple functionally independent energy networks as the first decoupling result.

[0094] Furthermore, the construction module 12 is used to perform the following steps: for the first decoupling result, divide the linear energy load and the nonlinear energy load; for each nonlinear energy load, perform linear approximation processing by introducing uncertain linear probabilities to determine the proxy linear energy load; based on the proxy linear energy load, replace the nonlinear energy load in the first decoupling result to determine the target energy structure.

[0095] Furthermore, the decision module 13 is used to perform the following steps: for the real-time campus energy scenario, determine the flexible load and the rigid load, wherein each flexible load is identified with energy slack; based on the flexible load and the rigid load, mark the target energy structure.

[0096] Furthermore, the decision module 13 is used to perform the following steps: setting a static energy consumption threshold and a dynamic energy consumption threshold on the load side for the park's energy consumption standards; determining the target energy structure based on the static energy consumption threshold and the dynamic energy consumption threshold, identifying the energy structure portion that exceeds the energy consumption limit as the energy structure exceeding the limit; and making an energy scheduling and balancing decision for the energy structure exceeding the limit.

[0097] Furthermore, the decision module 13 is used to perform the following steps: for the over-limit energy structure, identify rigid loads as the first priority, identify flexible loads and build an optimization space with energy relaxation; based on the first priority and the optimization space, with low energy consumption as the first objective and energy balance as the second objective, perform multi-objective optimization iterative decision-making to determine the energy management strategy, wherein the park microgrid is connected to the distribution network for coordinated energy supply.

[0098] Furthermore, the decision module 13 is used to perform the following steps: introduce a stability control database, wherein the stability control database contains multiple sequences representing fault scenarios-shutdown actions; perform energy operation risk control analysis based on the stability control database, and generate a stability control shutdown instruction; and incorporate the stability control shutdown instruction into the energy management strategy.

[0099] Furthermore, the decision module 13 is used to perform the following steps: based on the deployment and local recursive connection relationship of the intelligent monitoring device, the energy management strategy is decomposed to determine N sub-strategies, wherein the N sub-strategies correspond to the intelligent monitoring device; the N sub-strategies are transmitted in reverse to the intelligent monitoring device for distributed energy management guidance.

[0100] Through the foregoing detailed description of the park energy management method based on building energy consumption monitoring, those skilled in the art can clearly understand the park energy management method and system based on building energy consumption monitoring in this embodiment. As for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to the method section description.

[0101] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for park energy management based on building energy consumption monitoring, characterized in that, The method includes: The edge-end monitoring deployment is carried out, which involves the distributed deployment of intelligent monitoring devices with remote transmission capabilities to establish local recursive connections based on the park's local area and to connect to the energy management platform. By constructing a park simulation through park twins, and based on the focus of energy management, the park simulation is subjected to a first decoupling based on monitoring guidance and a second iteration based on linear approximation to determine the target energy structure. The monitoring guidance includes at least item-zone-household. The park simulation is updated with the network access data received by the platform. The first decoupling includes: when the focus is set as the sub-item dimension, determining the comprehensive energy network of the park simulation body, identifying the energy conversion nodes and energy transmission paths in the comprehensive energy network, generating multiple uncoupled energy subsystems based on the physical characteristics of energy flow, each subsystem retaining complete energy flow calculation attributes, and forming the first decoupling result from these functionally independent energy networks; Perform a second iteration based on linear approximation, including: Based on the first decoupling result, linear energy loads and nonlinear energy loads are divided; For each nonlinear energy load, an uncertain linear probability is introduced to perform linear approximation processing to determine the proxy linear energy load; Based on the proxy linear energy load, the nonlinear energy load in the first decoupling result is replaced to determine the target energy structure; For the target energy structure, energy consumption exceedance analysis and energy dispatch balance decision are performed to determine the energy management strategy and transmit it in reverse to the intelligent monitoring equipment for local management. Among them, the energy relaxation of the flexible load is used as the decision target, and stable operation constraint management is carried out by introducing power angle stabilization control and machine switching action.

2. The park energy management method based on building energy consumption monitoring as described in claim 1, characterized in that, After implementing edge-end monitoring deployment, the following is included: The intelligent monitoring device integrates monitoring data from locally recursively connected monitoring devices to determine the energy dataset, wherein each monitoring data is identified by a device spatiotemporal code; The energy dataset is uploaded to the network by the intelligent monitoring equipment and transmitted back to the energy management platform to update the park simulation.

3. The park energy management method based on building energy consumption monitoring as described in claim 1, characterized in that, Energy management focuses on at least one of the following: itemized, zoned, and household-based. Based on the energy management focus, the park simulation is decoupled; If energy management focuses on specific items, the decoupling method is as follows: The integrated energy network of the park simulation is determined, and the integrated energy network is decomposed according to multiple independent energy functions to determine multiple functionally independent energy networks as the first decoupling result.

4. The park energy management method based on building energy consumption monitoring as described in claim 1, characterized in that, Before performing energy consumption exceedance analysis and energy dispatch balance decision-making, the following should be included: For real-time energy scenarios in the park, flexible loads and rigid loads are identified, with each flexible load marked with its energy slack. The target energy structure is marked based on the flexible load and the rigid load.

5. The park energy management method based on building energy consumption monitoring as described in claim 4, characterized in that, Perform energy consumption exceedance analysis and energy dispatch balance decision-making, including: In accordance with the park's energy consumption standards, set static and dynamic energy consumption thresholds on the load side. The target energy structure is determined by the static energy consumption threshold and the dynamic energy consumption threshold, and the part of the energy structure that exceeds the energy consumption limit is identified as the energy structure that exceeds the limit. For the aforementioned energy structure that exceeds the limits, an energy dispatch and balancing decision is made.

6. The park energy management method based on building energy consumption monitoring as described in claim 5, characterized in that, For the aforementioned energy structure exceeding its limits, energy dispatch and balancing decisions are made, including: For the aforementioned over-limit energy structure, rigid loads are identified as the first priority, and flexible loads are identified and an optimization space is built based on energy relaxation. Based on the first priority and the optimization space, with low energy consumption as the first objective and energy balance as the second objective, a multi-objective optimization iterative decision is made to determine the energy management strategy, wherein the park microgrid is integrated into the distribution network for coordinated energy supply.

7. The park energy management method based on building energy consumption monitoring as described in claim 1, characterized in that, Stable operation control management is achieved by introducing power angle stabilization control for machine cutting actions, including: A stability control database is introduced, wherein the stability control database contains multiple sequences representing fault scenarios and machine switching actions; Based on the aforementioned stability control database, risk control analysis of energy operation is performed, and stability control shutdown instructions are generated; The stable control and machine switching command is incorporated into the energy management strategy.

8. The park energy management method based on building energy consumption monitoring as described in claim 1, characterized in that, Determine energy management strategies and transmit them in reverse to intelligent monitoring devices for local management, including: Based on the deployment and local recursive connection relationship of the intelligent monitoring equipment, the energy management strategy is decomposed to determine N sub-strategies, wherein the N sub-strategies correspond to the intelligent monitoring equipment; The N sub-strategies are transmitted in reverse to the intelligent monitoring device for distributed energy management guidance.

9. A park energy management system based on building energy consumption monitoring, characterized in that, The system is used to implement the park energy management method based on building energy consumption monitoring as described in any one of claims 1-8, the system comprising: The deployment module is used to perform edge-end monitoring deployment, which involves distributing intelligent monitoring devices with remote transmission capabilities to establish local recursive connections based on the park's local area and to connect to the energy management platform. The construction module is used to construct a park simulation by performing park twinning. Based on the energy management focus, the park simulation is subjected to a first decoupling based on monitoring guidance and a second iteration based on linear approximation to determine the target energy structure. The monitoring guidance includes at least item-zone-household. The park simulation is updated with the network access data received by the platform. The decision module is used to perform energy consumption over-limit analysis and energy dispatch balance decision for the target energy structure, determine the energy management strategy and transmit it in reverse to the intelligent monitoring equipment for local management. Among them, the energy relaxation of the flexible load is used as the decision target, and stable operation constraint management is carried out by introducing power angle stabilization control and machine switching action.

Images