A weak current system energy consumption optimization and capacity prediction method, system, device and medium based on digital twinning

Abstract

The application relates to a weak current system energy consumption optimization and capacity prediction method, system, device and medium based on digital twinning, and relates to the technical field of weak current equipment integration. The weak current system energy consumption optimization and capacity prediction method comprises the following steps: constructing a static twinning model according to the spatial position layout of a weak current system in a physical world and a weak current equipment connection relationship; converting energy consumption operation data of the weak current equipment, and transmitting and binding the energy consumption operation data to the static twinning model to generate a dynamic twinning model; generating an energy saving strategy suggestion list according to historical energy consumption data combined with preset energy saving rules, and optimizing the dynamic twinning model to obtain a dynamic energy saving twinning model; constructing a virtual operation instance according to basic operation data of a to-be-connected device, and deducing and evaluating a capacity use state combined with the dynamic energy saving twinning model; and generating a resource use evaluation report combined with preset different capacity thresholds; and the whole-process closed-loop intelligent management of the weak current system from digital modeling, energy consumption optimization to capacity prediction is realized.

Description

Technical Field

[0001] This application relates to the field of low-voltage equipment integration technology, and in particular to a method, system, device and medium for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins. Background Technology

[0002] With the rapid development of smart building and Internet of Things technologies, the current energy consumption optimization and capacity prediction of low-voltage systems mainly rely on real-time data monitoring based on SCADA systems, energy consumption records and statistical analysis of traditional time-series databases, and empirical prediction methods combined with simple regression models or grey theory. These methods have certain effects when the system scale is small and the load is stable.

[0003] Energy consumption of building low-voltage systems (such as network equipment, security hosts, and large display screens) is becoming increasingly significant, but current management is extensive and lacks refined management based on actual load.

[0004] Existing patents disclose a data center energy efficiency monitoring method based on digital twins, relating to the field of energy efficiency monitoring technology. This method addresses the problem of the inability to dynamically analyze the energy efficiency of data centers. The method includes: processing physical space data and real-time multi-source dynamic data of the data center; training a neural network based on historical multi-source dynamic data to obtain a digital twin prediction model; obtaining predicted power data of the data center based on real-time multi-source dynamic data, and analyzing the real-time equipment temperature and real-time ambient temperature of different devices within the data center; obtaining a standard range for real-time power data based on the predicted power data, and performing a comprehensive analysis of the real-time power data of the data center based on the standard range; and optimizing the connection methods between different devices within the data center using the digital twin prediction model. This invention achieves dynamic analysis of data center energy efficiency.

[0005] The existing technical solutions mentioned above have the following drawbacks: 1. When planning the expansion of weak current systems, they mainly rely on experience-based estimations, making it difficult to accurately predict the load pressure of new equipment on the existing network bandwidth, power supply capacity, and carrying platform, leading to over-investment or performance bottlenecks; 2. With the large-scale access of IoT devices, increasingly complex energy consumption behaviors, and the random access of new energy sources, static prediction models cannot adapt to dynamically changing equipment states, environmental factors, and human flow patterns. Each weak electronic system is independently controlled, lacking a multi-objective collaborative optimization strategy based on global energy efficiency. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the purpose of this application is to provide a method, system, equipment, and medium for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins. This enables large-scale, standardized, and rapid deployment and configuration migration of low-voltage equipment, greatly improving the efficiency of engineering implementation and operation and maintenance; ensuring the consistency and accuracy of parameter configuration; and reducing the risk of human error.

[0007] This was achieved using the following technical solutions: Firstly, this application provides a method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins, including: A static twin model is constructed based on the spatial layout of the low-voltage system in the physical world and the connection relationship of the low-voltage equipment. Collect and convert energy consumption operation data of low-voltage equipment, and transmit and bind it to a static twin model to generate a dynamic twin model; Based on historical energy consumption data and preset energy-saving rules, a list of energy-saving strategy suggestions is generated, and the dynamic twin model is optimized to obtain a dynamic energy-saving twin model. Based on the basic operating data of the devices to be connected, a virtual operating instance is constructed, and combined with a dynamic energy-saving twin model, the capacity usage status is deduced and evaluated. Based on different preset capacity thresholds, the system compares the capacity usage status and generates a resource usage assessment report.

[0008] By adopting the above technical solutions, a static twin model is constructed based on spatial topology and relational graph algorithms. A dynamic model is formed by binding real-time energy consumption data through IoT data fusion and time series alignment technology. Energy-saving strategies are generated with the help of prediction algorithms such as LSTM and rule engines. Then, simulation and load prediction algorithms are used to evaluate the capacity status. Finally, the digital simulation, optimization and early warning of energy consumption and capacity of weak current systems are realized, which significantly improves the level of energy efficiency management.

[0009] This application further specifies: based on the spatial layout of the low-voltage system in the physical world and the connection relationships of the low-voltage equipment, a static twin model is constructed, including: Semantic parsing and recognition are performed on the architectural drawings of buildings where low-voltage systems are deployed to obtain the floor structure; The spatial layout of the floor structure is transformed using lightweight methods to obtain a 3D building white model and floor plan. The floor plan is divided according to the functions of each area to form spatial objects, and metadata is bound to form equipment mounting containers. Based on the preset equipment deployment table, select the low-voltage equipment model and install it into the equipment mounting container according to the deployment location to form a virtual asset container; Based on the interaction logic of the low-voltage equipment, the virtual asset container is instance-associated to obtain the initial twin model; Based on the medium type and logical relationship of the connection between low-voltage equipment, cable generation and connectivity verification are performed on the initial twin model to construct a static twin model.

[0010] By adopting the above technical solution, the floor structure is extracted and a lightweight 3D white model is generated through semantic parsing algorithms of architectural drawings (CV and NLP). Based on the rule engine and metadata binding, the equipment mounting container is constructed, and graph theory algorithms are used to verify the connection relationship and generate the cable topology. This realizes the automated and high-precision construction of the static twin model of the low-voltage system, which significantly improves the modeling efficiency, data consistency and maintainability.

[0011] This application further specifies: collecting and converting energy consumption operation data of low-voltage equipment, transmitting and binding it to a static twin model, and generating a dynamic twin model, including: The data interface of the low-voltage equipment is parsed to extract connection parameters, point addresses and acquisition frequency, and a data acquisition device is constructed based on the interface type; The data acquisition device performs connection verification and data acquisition on the low-voltage equipment to obtain the original mixed data stream. The original mixed data stream is decoupled based on the device identifier to obtain single-device source data; The source data of a single device is parsed based on the data source identifier to extract the time-series runtime data; Illegal values ​​are filtered out and breakpoints are interpolated from the time-series data to obtain corrected time-series data. The modified time series data is normalized based on the static twin model to obtain standard time series data. A dynamic twin model is obtained by mapping virtual device objects and updating their real-time status on standard time-series data.

[0012] By adopting the above technical solution, a data collector is constructed based on protocol parsing and adaptive acquisition algorithms. The original time-series data is cleaned through anomaly detection and interpolation algorithms. The standard data is then bound to the twin model in real time using data normalization and object mapping algorithms. This achieves efficient and reliable acquisition of energy consumption data of weak current equipment and automatic generation of dynamic twin models, significantly improving data quality and model timeliness.

[0013] This application further specifies: performing virtual device object mapping and real-time state updates on standard time-series data to obtain a dynamic twin model, including: The standard timing data is statistically analyzed based on the device's runtime sequence, the number of data nodes is recorded, and the result is compared with a preset data trigger threshold. If the number of data nodes reaches the data trigger threshold, the current standard time-series data is received and distributed in combination with the data transmission rate to obtain device data packets; The device data packets are parsed and extracted to obtain the asset serial number and the device real-time data stream; Based on the asset serial number, the real-time data stream of the device is spatiotemporally aligned with the virtual device object to obtain a dynamic twin object; Based on the device type, the dynamic twin object is transformed into rules to obtain indicator status rules; Based on the status of the low-voltage equipment, conditional matching of the indicator status rules is performed to generate atomic rendering instructions; The virtual state object is obtained by incrementally updating and recalculating the state of the dynamic twin object according to the atomic rendering instructions. A dynamic twin model is obtained by globally instantiating and rendering the virtual state object.

[0014] By adopting the above technical solutions, real-time data streams are processed based on threshold triggering and dynamic distribution algorithms. Accurate mapping between data and twin objects is achieved through sequence number matching and spatiotemporal alignment technologies. Furthermore, a rule engine and incremental rendering algorithm are used to drive real-time updates of the twin model's state. This enables low-latency, high-fidelity synchronization and visualization of the dynamic twin model of the low-voltage system, significantly improving the real-time performance, interactivity, and decision support capabilities of the system monitoring.

[0015] This application further specifies: based on historical energy consumption data and preset energy-saving rules, a list of energy-saving strategy suggestions is generated, and the dynamic twin model is optimized to obtain a dynamic energy-saving twin model, including: Based on spatial calendar data, historical energy consumption time-series data and device metadata are spatiotemporally aligned to obtain regional device energy consumption data. Multi-dimensional hierarchical analysis of regional equipment energy consumption data yields time-series energy consumption curves, spatial energy density, and type-specific energy consumption characteristics. The preset energy-saving rules are analyzed from multiple dimensions to construct a list of energy-saving strategy suggestions, and time node factors, spatial node factors, and equipment type factors are extracted. Based on the time-series energy consumption curve and the time node factor, the time-series twin weight matrix of the dynamic twin model is modified to generate a time-series regular weight matrix. Based on the spatial energy consumption density and spatial node factors, the spatial twin weight matrix of the dynamic twin model is matched to generate a spatial regular weight matrix. Based on the energy consumption characteristics of different types and the equipment type factor, the type feature weight matrix of the dynamic twin model is fused to generate the type energy consumption weight matrix. The energy-saving rules are digitally converted, and the spatiotemporal fusion matrix and hierarchical equipment status matrix are extracted. Based on the spatiotemporal fusion matrix, the temporal rule weight matrix and the spatial rule weight matrix are weighted and iterated to obtain the spatiotemporal rule weight matrix. Based on the hierarchical device state matrix, the spatiotemporal rule weight matrix and the type energy consumption weight matrix are cross-trained to generate an energy-saving twin weight matrix; Based on the energy-saving twin weight matrix, historical energy consumption data is validated using examples, and the resource consumption and resource savings are output. If the difference between the resource consumption and the original consumption is within the preset tolerance range, then the difference between the resource savings and the original savings is calculated. If the difference ratio is within the tolerance range, then the current energy-saving rule is compared with the original energy-saving rule; If the two are the same, the current energy-saving twin weight matrix is ​​determined to be the core matrix of the dynamic energy-saving twin model; if the two are different, the current energy-saving rules are further transformed to generate energy-saving extension tags, and then fused and supplemented with the current energy-saving twin weight matrix to obtain the core matrix of the dynamic energy-saving twin model. If neither of them is within the corresponding tolerance range, then all parameters extracted by the current energy-saving rule are weighted and corrected in combination with the training constraint parameters to obtain the optimal energy-saving twin weight matrix as the core matrix of the dynamic energy-saving twin model.

[0016] By adopting the above technical solution, historical energy consumption data and equipment metadata are fused based on the spatiotemporal alignment algorithm. Temporal, spatial and type energy consumption characteristics are extracted through multidimensional analysis. Then, an optimized energy-saving twin weight matrix is ​​generated by using rule parsing and weight matrix iteration algorithms (such as spatiotemporal fusion weighting and cross-training). Finally, a dynamic energy-saving twin model that can accurately simulate and predict energy consumption is constructed, realizing intelligent diagnosis, strategy optimization and closed-loop verification of energy efficiency of weak current systems. This significantly improves the accuracy, adaptability and interpretability of energy-saving management.

[0017] This application further specifies: based on the basic operating data of the device to be connected, a virtual operating instance is constructed, and combined with a dynamic energy-saving twin model, the capacity usage status is deduced and evaluated, including: Based on the deployed device database, identify and match the devices to be accessed, and calculate the similarity of a single device. If the similarity of a single device is greater than the preset similarity threshold, the basic operating parameters of the currently deployed devices are extracted as the basic operating data of the devices to be connected. If not, then perform combination matching on the current devices to be connected, extract the basic operating parameters of the combined devices, and perform semantic weighted calculation in combination with the device function description to obtain the basic operating data of the current devices to be connected. Basic operational data is hierarchically encapsulated to generate a simulation parameter package, which is then integrated and associated with the virtual state object to generate a virtual operational object. Based on the power distribution logic diagram, hierarchical communication paths, and upper-layer service dependencies, the virtual running objects are topologically linked to construct virtual running instances; Based on the dynamic energy-saving twin model, a multi-dimensional logical deduction is performed on the virtual operation instance to calculate the capacity usage status; the capacity usage status includes power capacity, network bandwidth and load resources.

[0018] By adopting the above technical solution, the operating data of the device to be connected is adaptively obtained based on similarity matching and semantic weighting algorithms. Virtual operating instances are constructed through topological association, and multi-dimensional capacity status simulation evaluation is carried out by combining logical deduction model with dynamic energy-saving twin model. This enables accurate prediction and planning of power, bandwidth and load resources before the new device is connected, which significantly improves the expansion security and resource allocation foresight of the weak current system.

[0019] Secondly, this application also provides a digital twin-based energy consumption optimization and capacity prediction system for low-voltage systems, employing the following technical solution: A digital twin-based energy consumption optimization and capacity prediction system for low-voltage systems, comprising the following methods for implementing energy consumption optimization and capacity prediction in low-voltage systems: The static construction module is used to build a static twin model based on the spatial layout of the low-voltage system in the physical world and the connection relationship of the low-voltage equipment. The dynamic generation module is used to collect and convert energy consumption operation data of low-voltage equipment, and transmit and bind it to the static twin model to generate a dynamic twin model; The rule transformation module is used to optimize the dynamic twin model based on historical energy consumption data and preset energy-saving rules, so as to obtain a dynamic energy-saving twin model. The simulation and evaluation module is used to construct a virtual operation instance based on the basic operating data of the device to be connected, and combine it with a dynamic energy-saving twin model to simulate and evaluate the capacity usage status, compare it with different preset capacity thresholds, and generate a resource usage evaluation report.

[0020] By adopting the above technical solutions, a static twin model is constructed based on the spatial topology and relational graph algorithm. A dynamic energy-saving twin model is generated and optimized through IoT data fusion and LSTM prediction algorithm. The capacity status is evaluated by simulation and load prediction algorithm. This realizes the closed-loop intelligent management of the entire process of weak current system from digital modeling, energy consumption optimization to capacity prediction, which significantly improves the level of energy efficiency refinement and planning foresight.

[0021] Thirdly, this application also provides an electronic device, comprising: One or more processors; Memory, used to store one or more programs; When one or more programs are executed by one or more processors, the one or more processors implement any of the methods in the above scheme.

[0022] Fourthly, this application also provides a storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, at least one program, code set, or instruction set is loaded and executed by a processor to implement the energy consumption optimization and capacity prediction method for low-voltage systems based on digital twins as described above.

[0023] In summary, the beneficial technical effects of this application are as follows: A static twin model is constructed based on spatial topology and relational graph algorithms. A dynamic energy-saving twin model is generated and optimized through IoT data fusion and LSTM prediction algorithms. Simulation and load prediction algorithms are used to evaluate the capacity status, realizing closed-loop intelligent management of the entire process of weak current system from digital modeling, energy consumption optimization to capacity prediction, which significantly improves the level of energy efficiency refinement and planning foresight. Based on similarity matching and semantic weighting algorithms, the system adaptively acquires the operating data of the devices to be connected, constructs virtual operating instances through topological association, and uses a logical deduction model combined with a dynamic energy-saving twin model to conduct multi-dimensional capacity status simulation and evaluation. This enables accurate prediction and planning of power, bandwidth and load resources before the new devices are connected, improving the expansion security and resource allocation foresight of the low-voltage system. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the overall process of the energy consumption optimization and capacity prediction method for low-voltage systems in this application; Figure 2 This is a flowchart illustrating step S3 in the energy consumption optimization and capacity prediction method for low-voltage systems in this application. Figure 3 This is a schematic diagram of the structure of the low-voltage system energy consumption optimization and capacity prediction system in this application. Detailed Implementation

[0025] The present application will be further described in detail below with reference to the accompanying drawings.

[0026] Reference Figure 1 This application discloses a method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins, comprising: S1: Construct a static twin model based on the spatial layout of the low-voltage system in the physical world and the connection relationship of the low-voltage equipment; S2: Collect and convert energy consumption operation data of low-voltage equipment, and transmit and bind it to the static twin model to generate a dynamic twin model; S3: Based on historical energy consumption data and preset energy-saving rules, generate a list of energy-saving strategy suggestions and optimize the dynamic twin model to obtain a dynamic energy-saving twin model; S4: Based on the basic operating data of the devices to be connected, construct a virtual operating instance and combine it with a dynamic energy-saving twin model to deduce and evaluate the capacity usage status; S5: Based on different preset capacity thresholds, compare the capacity usage status and generate a resource usage assessment report.

[0027] In this embodiment, in the smart building management system of a high-rise office building, firstly, based on the building BIM model and the low-voltage floor plan, a static digital twin model is constructed for the equipment such as air conditioning terminal controllers, lighting circuits, and security sensors distributed on each floor according to their actual physical locations and RS-485 / Ethernet connection relationships; through smart meters and gateways deployed in the power distribution room, the energy consumption data of each low-voltage device (such as lighting power density and air conditioning fan frequency) is collected in real time, and after being converted by edge computing, it is dynamically bound to the twin model to form a dynamic twin reflecting the real-time operating status.

[0028] Further analysis of seasonal energy consumption curves and equipment start-up and shutdown logs over the past two years, combined with preset energy-saving rules (such as "automatically reducing area illuminance 30 minutes after personnel leave"), generates a suggestion list that includes time-domain optimization and equipment coordination strategies. Based on this, dynamic model parameters are optimized to construct a dynamic energy-saving twin model that can simulate the effects of the strategies.

[0029] When a building plans to add VRV low-voltage electrical equipment to an office area, the administrator inputs basic data such as the rated power and preset operating mode of the equipment into the system. The system then creates a corresponding virtual operating instance and performs Monte Carlo simulation based on a dynamic energy-saving twin model to assess the overall power capacity usage trend and peak load changes over the next year. Finally, the simulation results are automatically compared with preset "safety thresholds" (85%) and "early warning thresholds" (70%) to generate a resource usage assessment report that includes capacity margin analysis, recommended expansion time points, and potential overload risk warnings, providing decision support for the green operation and maintenance and scientific expansion of building infrastructure.

[0030] Preferably, step S1 includes: Semantic parsing and recognition are performed on the architectural drawings of buildings where low-voltage systems are deployed to obtain the floor structure; The spatial layout of the floor structure is transformed using lightweight methods to obtain a 3D building white model and floor plan. The floor plan is divided according to the functions of each area to form spatial objects, and metadata is bound to form equipment mounting containers. Based on the preset equipment deployment table, select the low-voltage equipment model and install it into the equipment mounting container according to the deployment location to form a virtual asset container; Based on the interaction logic of the low-voltage equipment, the virtual asset container is instance-associated to obtain the initial twin model; Based on the medium type and logical relationship of the connection between low-voltage equipment, cable generation and connectivity verification are performed on the initial twin model to construct a static twin model.

[0031] In this embodiment, the CAD drawings of the building are first imported, and the floors, walls and functional areas are automatically identified by the semantic parsing engine and converted into lightweight 3D white models and 2D floor plans. The system divides the space objects such as office areas, corridors and computer rooms according to the floor plans and binds metadata such as area and purpose to form equipment mounting containers. Then, according to the equipment deployment table, the corresponding 3D models of cameras, APs, sensors and other devices are automatically assembled into the designated positions of the containers to generate virtual asset containers.

[0032] Then, based on the control and communication logic between devices (such as the affiliation between cameras and NVRs), instance associations are established to form an initial twin model. Finally, based on the media types and topology logic such as "fiber optic" and "network cable" noted in the connection relationship table, the system automatically generates virtual cables in the model and performs connectivity verification, thereby constructing an accurate static twin model that fully maps to the physical building and includes all low-voltage equipment and their connection relationships.

[0033] Preferably, step S2 includes: A: Parse the data interface of the low-voltage equipment, extract the connection parameters, point addresses and acquisition frequency, and build a data acquisition device based on the interface type; B: The data acquisition device performs connection verification and data acquisition on the low-voltage equipment to obtain the raw mixed data stream; C: Decouple the original mixed data stream based on the device identifier to obtain single-device source data; D: Parse the source data of a single device based on the data source identifier and extract the time-series runtime data; E: Filter out illegal values ​​and interpolate breakpoints in the timing data to obtain corrected timing data; F: Normalize the modified time series data according to the static twin model to obtain standard time series data; G: Perform virtual device object mapping and real-time status updates on standard time-series data to obtain a dynamic twin model.

[0034] In this embodiment, in a high-rise office building where a static twin model has been constructed, the ModbusTCP and BACnet interface parameters of each low-voltage device (such as the air conditioning DDC controller and lighting sensor) are first parsed, and a data acquisition device is created accordingly to verify its connection status and collect the raw data stream.

[0035] Subsequently, the source data of a single device is parsed according to the device identifier, and time-series operating values ​​such as temperature and power consumption are extracted. After filtering out anomalies and linear interpolation, corrected data is obtained. Finally, based on the unique identifier of the device in the static model, the data is normalized and mapped to the corresponding virtual device object in real time, thereby dynamically updating the operating status and energy consumption curve of each device in the three-dimensional model, forming a dynamic twin model that truly reflects the real-time operating status of the building.

[0036] Preferably, step G includes: The standard timing data is statistically analyzed based on the device's runtime sequence, the number of data nodes is recorded, and the result is compared with a preset data trigger threshold. If the number of data nodes reaches the data trigger threshold, the current standard time-series data is received and distributed in combination with the data transmission rate to obtain device data packets; The device data packets are parsed and extracted to obtain the asset serial number and the device real-time data stream; Based on the asset serial number, the real-time data stream of the device is spatiotemporally aligned with the virtual device object to obtain a dynamic twin object; Based on the device type, the dynamic twin object is transformed into rules to obtain indicator status rules; Based on the status of the low-voltage equipment, conditional matching of the indicator status rules is performed to generate atomic rendering instructions; The virtual state object is obtained by incrementally updating and recalculating the state of the dynamic twin object according to the atomic rendering instructions. A dynamic twin model is obtained by globally instantiating and rendering the virtual state object.

[0037] In this embodiment, in the monitoring of low-voltage electrical equipment in the central office building, when the accumulated temperature and energy consumption time-series data of the air conditioning terminal equipment reaches the trigger threshold of 60 nodes per minute, the data packets for that period are received and distributed and parsed according to the transmission rate to extract the asset serial number and real-time data stream of each device; based on the serial number, the data is aligned with the virtual air conditioning equipment in the three-dimensional model in terms of timestamp and spatial location to form a dynamic twin object.

[0038] Subsequently, the data is transformed into indicator status rules such as "cooling efficiency" and "fan load rate" according to the type of air conditioning equipment. Atomized rendering instructions are generated based on preset status conditions (such as load rate > 85% being overload). According to the instructions, the colors and values ​​of the corresponding devices in the model are dynamically updated, and the temperature field distribution of the associated area is recalculated. Finally, a dynamic twin model with abnormal highlights and trend curves is rendered in real time in the twin platform, realizing global visual monitoring of the air conditioning operation status.

[0039] Preferably, refer to Figure 2 Step S3 includes: Based on spatial calendar data, historical energy consumption time-series data and device metadata are spatiotemporally aligned to obtain regional device energy consumption data. Multi-dimensional hierarchical analysis of regional equipment energy consumption data yields time-series energy consumption curves, spatial energy density, and type-specific energy consumption characteristics. The preset energy-saving rules are analyzed from multiple dimensions to construct a list of energy-saving strategy suggestions, and time node factors, spatial node factors, and equipment type factors are extracted. Based on the time-series energy consumption curve and the time node factor, the time-series twin weight matrix of the dynamic twin model is modified to generate a time-series regular weight matrix. Based on the spatial energy consumption density and spatial node factors, the spatial twin weight matrix of the dynamic twin model is matched to generate a spatial regular weight matrix. Based on the energy consumption characteristics of different types and the equipment type factor, the type feature weight matrix of the dynamic twin model is fused to generate the type energy consumption weight matrix. The energy-saving rules are digitally converted, and the spatiotemporal fusion matrix and hierarchical equipment status matrix are extracted. Based on the spatiotemporal fusion matrix, the temporal rule weight matrix and the spatial rule weight matrix are weighted and iterated to obtain the spatiotemporal rule weight matrix. Based on the hierarchical device state matrix, the spatiotemporal rule weight matrix and the type energy consumption weight matrix are cross-trained to generate an energy-saving twin weight matrix; Based on the energy-saving twin weight matrix, historical energy consumption data is validated using examples, and the resource consumption and resource savings are output. If the difference between the resource consumption and the original consumption is within the preset tolerance range, then the difference between the resource savings and the original savings is calculated. If the difference ratio is within the tolerance range, then the current energy-saving rule is compared with the original energy-saving rule; If the two are the same, the current energy-saving twin weight matrix is ​​determined to be the core matrix of the dynamic energy-saving twin model; if the two are different, the current energy-saving rules are further transformed to generate energy-saving extension tags, and then fused and supplemented with the current energy-saving twin weight matrix to obtain the core matrix of the dynamic energy-saving twin model. If neither of them is within the corresponding tolerance range, then all parameters extracted by the current energy-saving rule are weighted and corrected in combination with the training constraint parameters to obtain the optimal energy-saving twin weight matrix as the core matrix of the dynamic energy-saving twin model.

[0040] In this embodiment, when optimizing the energy efficiency of the central air conditioning system in a high-rise office building, the historical energy consumption time series data is first aligned with the metadata of equipment such as air conditioning units and fan coil units based on spatial calendar data (such as weekdays / holidays and office hours) to obtain the regional equipment energy consumption dataset for each floor. Then, through multi-dimensional hierarchical analysis, the daily peak and valley time series energy consumption curves, the spatial density of cooling capacity in each area, and the energy efficiency characteristics of different types of equipment are extracted.

[0041] Next, the preset energy-saving rules (such as "automatic adjustment of temperature control setpoint during non-office hours") are analyzed to construct a strategy suggestion list and extract time node factors (such as time period weight), spatial node factors (such as regional priority), and equipment type factors (such as equipment energy efficiency level). Based on the time-series energy consumption curve and time factors, the time-series weight matrix of the dynamic twin model is corrected to generate a time-series rule weight matrix. According to the spatial energy consumption density and spatial factors, a spatial rule weight matrix is ​​matched and generated. Combining type energy consumption characteristics and equipment type factors, a type energy consumption weight matrix is ​​generated. Then, the energy-saving rules are digitized, the spatiotemporal fusion matrix and the hierarchical equipment state matrix are extracted, and through weighted iteration and cross-training, the energy-saving twin weight matrix is ​​finally generated.

[0042] This matrix is ​​used to validate the energy consumption data from the past three months. The predicted resource consumption and resource savings are output and compared with the actual values ​​(with a preset tolerance range of ±5%). If the difference in consumption and the difference in savings are both within the tolerance range and the energy-saving rules remain unchanged, the weight matrix is ​​directly confirmed as the core matrix of the dynamic energy-saving twin model. If the rules are different, extended labels such as "time-sharing and zone-based dynamic temperature control" are added and the matrix is ​​fused and topologically supplemented to form the core matrix. If any proportion exceeds the tolerance range, all parameters of the rules are weighted and corrected in combination with the training constraint parameters to obtain the optimal weight matrix as the core matrix. Finally, a dynamic energy-saving twin model that can simulate, evaluate, and optimize the energy consumption of the air conditioning system is formed.

[0043] Preferably, step S4 includes: Based on the deployed device database, identify and match the devices to be accessed, and calculate the similarity of a single device. If the similarity of a single device is greater than the preset similarity threshold, the basic operating parameters of the currently deployed devices are extracted as the basic operating data of the devices to be connected. If not, then perform combination matching on the current devices to be connected, extract the basic operating parameters of the combined devices, and perform semantic weighted calculation in combination with the device function description to obtain the basic operating data of the current devices to be connected. Basic operational data is hierarchically encapsulated to generate a simulation parameter package, which is then integrated and associated with the virtual state object to generate a virtual operational object. Based on the power distribution logic diagram, hierarchical communication paths, and upper-layer service dependencies, the virtual running objects are topologically linked to construct virtual running instances; Based on the dynamic energy-saving twin model, a multi-dimensional logical deduction is performed on the virtual operation instance to calculate the capacity usage status; the capacity usage status includes power capacity, network bandwidth and load resources.

[0044] In this embodiment, the calculation logic for the power capacity projection is: future total power consumption of the circuit = current real-time total power consumption of the circuit + power consumption of the simulated equipment. This is compared with the circuit's rated load. The power distribution circuit is highlighted in the twin scenario. It is displayed in green / yellow / red according to the load rate (e.g., <80%, 80-95%, >95%). Specific values ​​are displayed: "Current load 1.2kW, new 0.05kW, total 1.25kW / 1.5kW".

[0045] The network bandwidth projection calculation logic is: future port bandwidth utilization = current real-time port traffic + estimated bandwidth from simulated devices. This is compared with the total port bandwidth. Simultaneously, the bandwidth impact on upstream links is recursively calculated. The target switch and key upstream links are highlighted. Affected ports are displayed as a bandwidth utilization bar chart. A notification states: "Core link utilization will increase to 78%."

[0046] The logic for calculating the load resources is: Future server load = Current server CPU / memory utilization + Estimated load from simulated devices. This is compared to the server's performance thresholds. Associated server models are highlighted. The predicted CPU / memory utilization after the increase in load is displayed in a chart format.

[0047] In this embodiment, when expanding the security system of a high-rise office building and planning to add a batch of facial recognition access control terminals, the system first matches them with the deployed device library and calculates the similarity of a single device (the preset similarity threshold range is 88%-92%, determined based on the matching weight of device model, chip platform, and interface protocol). If the device is identified as a device from the same series and the similarity is greater than 90%, the rated power, peak network traffic, and other parameters of the existing terminals in the library are directly used as its basic operating data. If the device is a new model (similarity less than 88%), the estimated operating data is generated by combining matching parameters (such as referring to the mixed parameters of access control and card readers of the same brand) and semantically weighting calculation based on its "binocular liveness detection gate" function description.

[0048] Subsequently, this data is encapsulated into a simulation parameter package containing electrical, network, and computing layers, and merged with the corresponding virtual devices in the twin model to generate a virtual running object. Next, based on the branch capacity of the power distribution cabinet, the remaining bandwidth of the PoE switch, and the current API load of the face analysis server, a complete topological association of the object is established to form an interactive virtual running instance. Finally, based on the dynamic energy-saving twin model, Monte Carlo simulation is performed on this instance to accurately calculate the real-time and peak usage status of the added equipment on the floor circuit (power capacity), backbone network (bandwidth occupancy), and cloud services (GPU load resources), providing quantitative decision support for the feasibility of expansion.

[0049] Preferably, step S5 includes: List all affected system components in a structured format, including power capacity, network bandwidth, and load resources, clearly marking their "safe," "warning," or "over-limit" status. Provide key data such as: total load after addition, remaining safety margin, and location of risk points.

[0050] If the simulation results show warnings or exceed limits, intelligent suggestions will be provided: Power supply: "It is recommended to connect the device to 'Circuit B' with a larger load margin (currently 300W)." Network: "The target switch port bandwidth is insufficient. It is recommended to switch to Gigabit port Gig1 / 0 / 5 (current utilization is 20%)." Deployment: "The authentication server is already under high load. It is recommended to deploy a new authentication service instance on a backup server." Based on the suggestions, adjust the parameters or location of the simulation equipment to create multiple "simulation schemes".

[0051] Multiple solutions are simulated and compared in parallel to select the optimal solution.

[0052] Once the simulation plan is confirmed to be feasible, a standardized deployment work order containing equipment model, installation location, network configuration, and power supply requirements is automatically generated and directly dispatched to the operation and maintenance or construction team.

[0053] Based on the network parameters (VLAN, IP, etc.) determined in the simulation, the system can generate device initialization configuration scripts in advance, improving the efficiency of going online.

[0054] Once the scheme is confirmed, the simulation instance can be switched to the "planning" state, and its estimated load can be incorporated into the future capacity baseline calculation to achieve forward-looking capacity planning.

[0055] Reference Figure 3 A digital twin-based energy consumption optimization and capacity prediction system for low-voltage systems, applied to the energy consumption optimization and capacity prediction method for low-voltage systems, includes: The static construction module is used to build a static twin model based on the spatial layout of the low-voltage system in the physical world and the connection relationship of the low-voltage equipment. The dynamic generation module is used to collect and convert energy consumption operation data of low-voltage equipment, and transmit and bind it to the static twin model to generate a dynamic twin model; The rule transformation module is used to optimize the dynamic twin model based on historical energy consumption data and preset energy-saving rules, so as to obtain a dynamic energy-saving twin model. The simulation and evaluation module is used to construct a virtual operation instance based on the basic operating data of the device to be connected, and combine it with a dynamic energy-saving twin model to simulate and evaluate the capacity usage status, compare it with different preset capacity thresholds, and generate a resource usage evaluation report.

[0056] The implementation principle of this embodiment is as follows: The static construction module first imports the building's BIM drawings and low-voltage point location table. Based on the actual physical location of the equipment and the network topology, it constructs an accurate static digital twin model containing equipment such as air conditioning DDC controllers, lighting circuit sensors, and security cameras. Subsequently, the dynamic generation module collects the operating energy consumption data of each low-voltage device in real time through smart meters and data acquisition gateways deployed at the site. After edge computing conversion, the data is bound to the corresponding virtual devices in the static model, forming a three-dimensional dynamic twin model that reflects the system status in real time. On this basis, the rule conversion module uses historical energy consumption curves and equipment operation logs from the past two years, combined with preset energy-saving rules (such as "fresh air linkage adjustment based on personnel density"), to optimize the parameters and behavioral logic of the dynamic model using algorithms, generating a dynamic energy-saving twin model that can simulate and evaluate the effect of energy-saving strategies.

[0057] When a building management company plans to add a VRV air conditioning and intelligent lighting system to an office area, the simulation and evaluation module encapsulates the basic operating parameters of the new equipment (such as rated power and communication protocol) into a virtual operating instance and imports it into a dynamic energy-saving twin model. Through Monte Carlo simulation, it simulates the occupancy status of the building's total power capacity, local area network bandwidth, and central server load under different seasonal conditions in the coming year. The results are then automatically compared with preset multi-level capacity thresholds such as "safety," "early warning," and "critical," ultimately generating a detailed resource usage assessment report that provides a quantitative basis for infrastructure expansion and green operation and maintenance.

[0058] An electronic device, comprising: One or more processors; Memory, used to store one or more programs; When one or more programs are executed by one or more processors, the one or more processors implement any of the methods in the above scheme.

[0059] A storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, at least one program, code set, or instruction set is loaded and executed by a processor to implement the energy consumption optimization and capacity prediction method for low-voltage systems as described above.

[0060] The embodiments described in this specific implementation are preferred embodiments of this application and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins, characterized in that, include: A static twin model is constructed based on the spatial layout of the low-voltage system in the physical world and the connection relationship of the low-voltage equipment. Collect and convert energy consumption operation data of low-voltage equipment, and transmit and bind it to the static twin model to generate a dynamic twin model; Based on historical energy consumption data and preset energy-saving rules, a list of energy-saving strategy suggestions is generated, and the dynamic twin model is optimized to obtain a dynamic energy-saving twin model. Based on the basic operating data of the devices to be connected, a virtual operating instance is constructed, and combined with the dynamic energy-saving twin model, the capacity usage status is deduced and evaluated. Based on different preset capacity thresholds, the capacity usage status is compared to generate a resource usage assessment report.

2. The method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins according to claim 1, characterized in that, The static twin model is constructed based on the spatial layout of the low-voltage system in the physical world and the connection relationships of the low-voltage equipment, including: Semantic parsing and recognition are performed on the architectural drawings of buildings where low-voltage systems are deployed to obtain the floor structure; The spatial layout of the floor structure is subjected to lightweight transformation to obtain a three-dimensional building white model and floor plan; The floor plan is divided according to the regional functions to form spatial objects, and metadata is bound to form equipment mounting containers. The low-voltage equipment model is selected according to the preset equipment deployment table, and installed into the equipment mounting container according to the deployment location to form a virtual asset container; Based on the interaction logic of the low-voltage equipment, the virtual asset container is instance-associated to obtain an initial twin model; Based on the medium type and logical relationship of the connection between low-voltage equipment, cable generation and connectivity verification are performed on the initial twin model to construct a static twin model.

3. The method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins according to claim 1, characterized in that, The process of collecting and converting the energy consumption operation data of the low-voltage equipment, transmitting and binding it to the static twin model, and generating a dynamic twin model includes: The data interface of the low-voltage equipment is parsed to extract connection parameters, point addresses and acquisition frequency, and a data acquisition device is constructed based on the interface type; The data acquisition device is used to perform connection verification and data acquisition on the low-voltage equipment to obtain the original mixed data stream. The original mixed data stream is decoupled based on the device identifier to obtain single-device source data; The single-device source data is parsed based on the data source identifier to extract time-series runtime data; Illegal values ​​are filtered out and breakpoints are interpolated from the time-series running data to obtain corrected time-series data; The modified time series data is normalized according to the static twin model to obtain standard time series data. The standard time-series data is mapped to virtual device objects and updated in real time to obtain a dynamic twin model.

4. The method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins according to claim 3, characterized in that, The process of mapping virtual device objects and updating their real-time status on the standard time-series data to obtain a dynamic twin model includes: The standard timing data is statistically analyzed based on the device's runtime sequence, the number of data nodes is recorded, and the result is compared with a preset data trigger threshold. If the number of data nodes reaches the data trigger threshold, the current standard time-series data is received and distributed in combination with the data transmission rate to obtain a device data packet; The device data packets are parsed and extracted to obtain the asset serial number and the device real-time data stream; Based on the asset serial number, the real-time data stream of the device is spatiotemporally aligned with the virtual device object to obtain a dynamic twin object; The dynamic twin object is transformed according to the device type to obtain the indicator status rules; Based on the status of the low-voltage equipment, the indicator status rules are conditionally matched to generate atomic rendering instructions. The dynamic twin object is incrementally updated and its state is recalculated according to the atomic rendering instructions to obtain a virtual state object. The virtual state object is globally instantiated and rendered to obtain a dynamic twin model.

5. The method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins according to claim 1, characterized in that, The process involves generating a list of energy-saving strategy suggestions based on historical energy consumption data and preset energy-saving rules, and optimizing the dynamic twin model to obtain a dynamic energy-saving twin model, including: Based on spatial calendar data, historical energy consumption time-series data and device metadata are spatiotemporally aligned to obtain regional device energy consumption data. Multi-dimensional hierarchical analysis was performed on the energy consumption data of the equipment in the region to obtain time-series energy consumption curves, spatial energy consumption density, and type energy consumption characteristics. The preset energy-saving rules are analyzed from multiple dimensions to construct a list of energy-saving strategy suggestions, and time node factors, spatial node factors, and equipment type factors are extracted. Based on the time-series energy consumption curve and the time node factor, the time-series twin weight matrix of the dynamic twin model is modified to generate a time-series regular weight matrix. Based on the spatial energy consumption density and the spatial node factor, the spatial twin key matrix of the dynamic twin model is matched to generate a spatial rule weight matrix. Based on the energy consumption characteristics of the type and the equipment type factor, the type feature weight matrix of the dynamic twin model is fused to generate the type energy consumption weight matrix.

6. The method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins according to claim 5, characterized in that, The process of generating an energy-saving strategy suggestion list based on historical energy consumption data and preset energy-saving rules, and optimizing the dynamic twin model to obtain a dynamic energy-saving twin model, further includes: The energy-saving rules are digitally converted, and the spatiotemporal fusion matrix and hierarchical equipment status matrix are extracted. Based on the spatiotemporal fusion matrix, the temporal rule weight matrix and the spatial rule weight matrix are weighted and iterated to obtain the spatiotemporal rule weight matrix. Based on the hierarchical device state matrix, the spatiotemporal rule weight matrix and the type energy consumption weight matrix are cross-trained to generate an energy-saving twin weight matrix; Based on the energy-saving twin weight matrix, historical energy consumption data is validated using examples, and the resource consumption and resource savings are output. If the difference between the resource consumption and the original consumption is within a preset tolerance range, then the difference between the resource savings and the original savings is calculated. If the phase difference ratio is within the tolerance range, then the current energy-saving rule is compared with the original energy-saving rule; If the two are the same, the current energy-saving twin weight matrix is ​​determined to be the core matrix of the dynamic energy-saving twin model; if the two are different, the current energy-saving rules are further transformed to generate energy-saving extension tags, and then fused and supplemented with the current energy-saving twin weight matrix to obtain the core matrix of the dynamic energy-saving twin model. If neither of them is within the corresponding tolerance range, then all parameters extracted by the current energy-saving rule are weighted and corrected in combination with the training constraint parameters to obtain the optimal energy-saving twin weight matrix as the core matrix of the dynamic energy-saving twin model.

7. The method for energy consumption optimization and capacity prediction of low-voltage systems based on digital twins according to claim 1, characterized in that, The process of constructing a virtual operating instance based on the basic operating data of the device to be connected, and combining it with the dynamic energy-saving twin model to deduce and evaluate the capacity usage status, includes: Based on the deployed device database, identify and match the devices to be accessed, and calculate the similarity of a single device. If the similarity of a single device is greater than a preset similarity threshold, then the basic operating parameters of the currently deployed devices are extracted as the basic operating data of the devices to be connected. If not, then perform combination matching on the current devices to be connected, extract the basic operating parameters of the combined devices, and perform semantic weighted calculation in combination with the device function description to obtain the basic operating data of the current devices to be connected. The basic operating data is hierarchically encapsulated to generate a simulation parameter package, which is then fused and associated with the virtual state object to generate a virtual operating object; Based on the power distribution logic diagram, hierarchical communication paths, and upper-layer service dependencies, the virtual running object is topologically correlated to construct a virtual running instance; The virtual operating instance is subjected to multi-dimensional logical deduction based on the dynamic energy-saving twin model to calculate the capacity usage status, which includes power capacity, network bandwidth, and load resources.

8. A digital twin-based energy consumption optimization and capacity prediction system for low-voltage systems, used to implement the energy consumption optimization and capacity prediction method for low-voltage systems as described in any one of claims 1-7, characterized in that, include: The static construction module is used to build a static twin model based on the spatial layout of the low-voltage system in the physical world and the connection relationship of the low-voltage equipment. The dynamic generation module is used to collect and convert the energy consumption operation data of the low-voltage equipment, and transmit and bind it to the static twin model to generate the dynamic twin model; The rule transformation module is used to optimize the dynamic twin model based on historical energy consumption data and preset energy-saving rules to obtain a dynamic energy-saving twin model. The simulation and evaluation module is used to construct a virtual operation instance based on the basic operating data of the device to be connected, and combine it with the dynamic energy-saving twin model to simulate and evaluate the capacity usage status, compare it with different preset capacity thresholds, and generate a resource usage evaluation report.

9. An electronic device, comprising: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the energy consumption optimization and capacity prediction method for low-voltage systems as described in any one of claims 1 to 7.

10. A storage medium storing at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the energy consumption optimization and capacity prediction method for a low-voltage system as described in any one of claims 1 to 7.

Images