A digital twin energy efficiency assessment method and system for green building renovation

Abstract

This invention discloses a digital twin energy efficiency assessment method and system for green building retrofits, belonging to the field of digital twin technology. The method includes: acquiring measured operational and meteorological sequences to construct a directed energy transfer graph and extracting a dynamic baseline energy consumption feature set; mapping retrofit parameters to generate a digital twin simulation map, and deriving and extracting energy efficiency improvement predictions and load time-series offset sequences; substituting the above feature sequences and economic and climate disturbance matrices into a Markov chain Monte Carlo model for sampling evolution, and analyzing and extracting the cost-effectiveness probability distribution and system flexible response boundary; constructing a multidimensional carbon economic evaluation feature set, performing non-dominated ranking optimization, and outputting an optimal decision map that satisfies commercial risk. The embodiments of this application effectively quantify the commercial investment risk under market fluctuations and significantly improve the economic efficiency and reliability of retrofit decisions.

Description

Technical Field

[0001] This invention relates to the field of digital twin technology, specifically to a digital twin energy efficiency assessment method and system for green building renovation. Background Technology

[0002] With the comprehensive advancement of green building renovation projects, the digital management and scientific planning of building energy consumption have gradually become the core of industry development. As the main energy consumer, the economic and energy-saving assessment of building electromechanical systems' renovation plans directly affects the project's effectiveness. Digital twin technology, by constructing virtual mapping models of physical entities, provides a new perspective and calculation method for the energy efficiency assessment of complex systems. Currently, many existing buildings undergoing low-carbon upgrades urgently need scientific assessment methods to guide equipment replacement and the optimization of control strategies, ensuring that project investments achieve the expected energy-saving returns.

[0003] Existing energy efficiency assessment methods primarily employ static simulation software to calculate the energy consumption of individual devices. This approach isolates the operational states of each electromechanical system, failing to comprehensively consider the thermoelectric coupling effects between systems and the nonlinear energy efficiency degradation caused by equipment aging. This results in significant discrepancies between the calculated results and actual operating conditions. Furthermore, existing assessment models rely solely on fixed energy prices and ideal external weather conditions for revenue prediction, failing to effectively address the dynamic economic risks arising from frequent extreme weather events and fluctuations in carbon emission trading market prices. Consequently, the resulting retrofit decisions are highly susceptible to significantly extended investment payback periods or even severe losses in practical engineering applications. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a digital twin energy efficiency assessment method and system for green building renovation, which solves the problems mentioned in the background.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a digital twin energy efficiency assessment method for green building renovation, comprising the following steps: S1, acquiring the measured operation sequence and meteorological sequence of the building to be renovated, and constructing a directed energy transfer graph representing the electromechanical supply and demand logic; extracting the energy efficiency decay rate and heat transfer delay of the electromechanical nodes in the directed energy transfer graph, and constructing a dynamic baseline energy consumption feature set; S2, mapping the preset renovation scheme parameters into the directed energy transfer graph for reconstruction, generating a digital twin simulation map; performing multi-device coupled load extrapolation based on the digital twin simulation map, and extracting the predicted energy efficiency improvement sequence of the electromechanical nodes. S3. Obtain the economic disturbance matrix containing carbon electricity price fluctuations and the climate disturbance matrix containing extreme weather probabilities. Substitute the dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time series offset sequence, economic disturbance matrix and climate disturbance matrix into the Markov chain Monte Carlo model for sampling evolution, and analyze and extract the cost-effectiveness probability distribution set and system flexible response boundary. S4. Construct a multidimensional carbon economic evaluation feature set based on the cost-effectiveness probability distribution set and system flexible response boundary. Perform non-dominated ranking optimization on the multidimensional carbon economic evaluation feature set, and output the optimal transformation decision map that meets the preset business risk threshold.

[0006] Furthermore, the specific process of acquiring the measured operation sequence and meteorological sequence of the building to be renovated, and constructing a directed graph representing the electromechanical supply and demand logic, is as follows: Collect multi-dimensional operating status parameters of the chiller units, water pumps, and air conditioning terminals at the bottom layer of the IoT system of the building to be renovated; simultaneously collect outdoor temperature and humidity parameters from the external meteorological monitoring terminal; perform multi-source time series timestamp alignment to generate the measured operation sequence and meteorological sequence of the building to be renovated; analyze the electromechanical-hydraulic connection topology logic of the chiller units, water pumps, and air conditioning terminals within the measured operation sequence of the building to be renovated; establish nodes and weighted edges that map the electromechanical-hydraulic connection topology logic; integrate nodes, weighted edges, the measured operation sequence of the building to be renovated, and the meteorological sequence to perform graph mapping structure assembly to generate a directed graph representing the electromechanical supply and demand logic.

[0007] Furthermore, the specific process of extracting the energy efficiency decay rate and heat transfer delay of electromechanical nodes in the directed energy transfer graph and constructing a dynamic baseline energy consumption feature set is as follows: Track the deviation between the operating state of the energy efficiency benchmark data on the nameplate of the nodes in the directed energy transfer graph and the measured load response parameters, and analyze the energy efficiency decay rate of the electromechanical nodes; perform a time-series traversal of the cross-correlation function on the input thermal peak data and output thermal valley data of adjacent nodes in the directed energy transfer graph, locate the waveform alignment corresponding time delay interval, and extract the heat transfer delay across nodes; splice the energy efficiency decay rate, heat transfer delay and the measured operating sequence of the building to be modified and perform tensor fusion to generate a dynamic baseline energy consumption feature set.

[0008] Furthermore, the specific process of mapping the preset modification scheme parameters into the directed energy transfer graph for reconstruction and generating a digital twin simulation graph is as follows: Analyze the new host performance characteristics, pipeline resistance variation characteristics, and control scheduling variation strategies embedded in the preset modification scheme parameters; unload high-energy-consumption aging nodes from the directed energy transfer graph, implant new nodes corresponding to the new host performance characteristics, and refresh the weighted edge connectivity resistance matrix of the directed energy transfer graph according to the pipeline resistance variation characteristics; transform the control scheduling variation strategy into node collaborative scheduling constraints, reconstruct the topological mapping relationship of the directed energy transfer graph, and generate a digital twin simulation graph.

[0009] Furthermore, based on the digital twin simulation map, the specific process of performing multi-device coupled load extrapolation and extracting the energy efficiency improvement prediction sequence and load time series offset sequence of electromechanical nodes is as follows: Apply historical full-condition heat load boundary conditions to the digital twin simulation map, trigger the whole network heat and power coordinated cascade response along the weighted edge direction, and perform multi-device coupled load extrapolation; collect the dynamic energy efficiency parameters of the nodes generated by the multi-device coupled load extrapolation, map and compare them with the corresponding nodes in the dynamic baseline energy consumption feature set to extract the gain amplitude features, and establish the energy efficiency improvement prediction sequence of electromechanical nodes; track the timing span characteristics of the cold storage and release of nodes triggered by the control and scheduling change strategy, analyze the trajectory of the power load peak on the time axis, and construct the load time series offset sequence.

[0010] Furthermore, the specific process for obtaining the economic disturbance matrix containing carbon electricity price fluctuations and the climate disturbance matrix containing extreme weather probabilities is as follows: Collect historical settlement price time-series data from the carbon emission trading market and time-of-use electricity price transition time-series data from the regional power grid; analyze the frequency domain characteristics of price fluctuations and the confidence interval boundaries to generate the economic disturbance matrix; obtain historical extreme high and low temperature frequency samples corresponding to the geographical coordinates of the building to be renovated; fit the probability density function of extreme temperature and humidity meteorological elements; extract discrete meteorological mutation vectors exceeding the baseline climate conditions to generate the climate disturbance matrix; align the time step of price evolution within the economic disturbance matrix with the execution timestamp of mutation occurrence time step within the climate disturbance matrix, and output the economic disturbance matrix and the climate disturbance matrix.

[0011] Furthermore, the specific process of substituting the dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time-series offset sequence, economic disturbance matrix, and climate disturbance matrix into the Markov chain Monte Carlo model sampling evolution, and extracting the cost-effectiveness probability distribution set and the system flexible response boundary is as follows: The dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, and load time-series offset sequence are mapped to the initial state space of the Markov chain Monte Carlo model; the economic disturbance matrix and climate disturbance matrix are transformed into random walk state transition probability parameters of the Markov chain Monte Carlo model; the Markov chain Monte Carlo model is triggered to perform a cross-lifecycle time-step random sampling walk; the initial state space and state transition probability parameters are fused to perform dimensionality reduction of carbon assets and energy consumption input features; the net present value evolution trajectory under the clustering convergence state is clustered to generate the cost-effectiveness probability distribution set; the extreme values ​​of the load time-series offset response are tracked under the limit state containing discrete meteorological mutation vectors during the random sampling walk; the maximum load peak-shaving and valley-filling capacity of electromechanical network nodes under the condition of not violating thermal comfort constraints is analyzed; and the system flexible response boundary is extracted.

[0012] Furthermore, based on the cost-effectiveness probability distribution set and the system flexible response boundary, a multidimensional carbon economy evaluation feature set is constructed. The specific process of performing non-dominated ranking optimization on this feature set to output the optimal retrofit decision graph that satisfies the preset business risk threshold is as follows: The cost-effectiveness probability distribution set is dimensionality-reduced and mapped to an expected investment return feature vector; the system flexible response boundary is transformed into a grid interaction elasticity feature vector; the expected investment return feature vector and the grid interaction elasticity feature vector are fused to generate the multidimensional carbon economy evaluation feature set; a preset business risk threshold is set as the optimization constraint boundary; the multidimensional carbon economy evaluation feature set is substituted into the non-dominated ranking algorithm space to perform Pareto front dominance hierarchical filtering, extracting the optimal Pareto front solution set that fits the optimization constraint boundary; the corresponding retrofit equipment combinations and control logic within the optimal Pareto front solution set are mapped to a visualized node relationship network, generating the optimal retrofit decision graph.

[0013] A digital twin energy efficiency assessment system for green building renovation, used to execute the aforementioned digital twin energy efficiency assessment method for green building renovation, includes: a baseline extraction module for acquiring the measured operating sequence and meteorological sequence of the building to be renovated, and constructing a directed energy transfer graph representing the electromechanical supply and demand logic; extracting the energy efficiency decay rate and heat transfer delay of the electromechanical nodes in the directed energy transfer graph, and constructing a dynamic baseline energy consumption feature set; a twin simulation module for mapping the preset renovation scheme parameters into the directed energy transfer graph for reconstruction, generating a digital twin simulation map; and performing multi-device coupled load simulation based on the digital twin simulation map to extract the predicted energy efficiency improvement sequence of the electromechanical nodes. The system includes a load time-series offset sequence; a sampling evolution module, used to obtain the economic disturbance matrix containing carbon electricity price fluctuations and the climate disturbance matrix containing extreme weather probabilities; a dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time-series offset sequence, economic disturbance matrix and climate disturbance matrix are substituted into the Markov chain Monte Carlo model for sampling evolution, and the cost-effectiveness probability distribution set and system flexible response boundary are extracted analytically; a decision optimization module is used to construct a multidimensional carbon economic evaluation feature set based on the cost-effectiveness probability distribution set and system flexible response boundary, perform non-dominated ranking optimization on the multidimensional carbon economic evaluation feature set, and output the optimal transformation decision map that meets the preset business risk threshold.

[0014] The present invention has the following beneficial effects: (1) A digital twin energy efficiency assessment method for green building renovation, which obtains the measured operation sequence and meteorological sequence of the building to be renovated, constructs a directed energy transfer graph representing the electromechanical supply and demand logic, extracts the energy efficiency decay rate and heat transfer delay of the electromechanical nodes in the directed energy transfer graph, constructs a dynamic baseline energy consumption feature set, and then maps the preset renovation scheme parameters into the directed energy transfer graph for reconstruction, generating a digital twin simulation map. Based on the digital twin simulation map, multi-equipment coupled load extrapolation is performed, and the energy efficiency improvement prediction sequence and load time series offset sequence of the electromechanical nodes are extracted. The above steps transform the physical state of the building's actual operation into a quantifiable directed energy transfer graph, effectively extracting the actual decay of old equipment and the hysteresis characteristics of the pipeline network. At the same time, the thermo-electric coupling cascade effect between multiple electromechanical systems after the introduction of new equipment is extrapolated in the digital twin simulation map, overcoming the defect of the existing method of calculating equipment in isolation, and significantly improving the physical reality and calculation accuracy of energy efficiency assessment.

[0015] (2) A digital twin energy efficiency assessment system for green building renovation, which obtains an economic disturbance matrix containing carbon electricity price fluctuations and a climate disturbance matrix containing extreme weather probabilities, substitutes various characteristic parameters into a Markov chain Monte Carlo model for sampling evolution, analyzes and extracts the cost-effectiveness probability distribution set and the system flexible response boundary, and finally constructs a multidimensional carbon economic evaluation feature set based on the above probability distribution set and flexible response boundary, performs non-dominated ranking optimization, and outputs the optimal renovation decision map that meets the preset business risk threshold. The above steps take into account dynamic disturbance factors such as external climate change and market price fluctuations, quantify the risk resistance of the renovation scheme in a real complex environment using sampling evolution, and comprehensively optimize and analyze multidimensional economic indicators, avoiding the investment loss risk faced by traditional static economic calculations in actual operation, and providing a scientific and reliable decision map for renovation projects.

[0016] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0017] Figure 1 This is a flowchart of a digital twin energy efficiency assessment method for green building renovation according to the present invention; Figure 2 This is a schematic diagram comparing the time-series load shifts under all operating conditions before and after the modification, based on the digital twin simulation graph in this embodiment of the application. Figure 3 This is a schematic diagram of the full lifecycle cost-effectiveness probability distribution based on the sampling evolution of the Markov chain Monte Carlo model in the embodiments of this application; Figure 4 This is a schematic diagram of Pareto frontier optimization for multidimensional carbon economy evaluation in the embodiments of this application; Figure 5 This is a flowchart of a digital twin energy efficiency assessment system for green building renovation according to the present invention. Detailed Implementation

[0018] This application provides a digital twin energy efficiency assessment method and system for green building renovation, which solves the problem that existing energy efficiency assessment methods calculate equipment energy consumption in isolation and cannot cope with the risks of dynamic environmental and economic fluctuations.

[0019] The overall approach of the scheme in this application is as follows: First, a directed energy transfer graph is constructed based on the measured sequence to accurately depict the dynamic energy consumption base of the existing building. Then, the parameters of the renovation scheme are mapped onto the directed graph to generate a digital twin simulation map and deduce the load evolution law of multi-device coupling. Next, a dual perturbation matrix of climate and economy is introduced to perform multi-dimensional sampling evolution to extract the risk-return probability distribution of the entire life cycle. Finally, an optimal renovation decision map that takes into account both energy-saving effect and risk resistance capability is generated through multi-objective optimization analysis, thereby realizing a scientific and comprehensive commercial and technical comprehensive evaluation of the green building renovation scheme.

[0020] Example 1; please refer to Figure 1 This invention provides a technical solution: a digital twin energy efficiency assessment method for green building renovation, comprising the following steps: S1, acquiring the measured operation sequence and meteorological sequence of the building to be renovated, and constructing a directed energy transfer graph representing the electromechanical supply and demand logic; extracting the energy efficiency decay rate and heat transfer delay of the electromechanical nodes in the directed energy transfer graph, and constructing a dynamic baseline energy consumption feature set; S2, mapping the preset renovation scheme parameters into the directed energy transfer graph for reconstruction, generating a digital twin simulation map; performing multi-device coupled load extrapolation based on the digital twin simulation map, and extracting the energy efficiency improvement prediction sequence and load of the electromechanical nodes. S3. Obtain the economic disturbance matrix containing carbon electricity price fluctuations and the climate disturbance matrix containing extreme weather probabilities. Substitute the dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time-series offset sequence, economic disturbance matrix and climate disturbance matrix into the Markov chain Monte Carlo model for sampling evolution, and analyze and extract the cost-effectiveness probability distribution set and system flexible response boundary. S4. Construct a multidimensional carbon economic evaluation feature set based on the cost-effectiveness probability distribution set and system flexible response boundary. Perform non-dominated ranking optimization on the multidimensional carbon economic evaluation feature set, and output the optimal transformation decision map that meets the preset business risk threshold.

[0021] In this implementation plan, step S1 is primarily used to establish the actual energy consumption baseline state of the existing building. Specifically, the system first synchronously acquires historical time-series data generated by the actual operation of IoT devices inside the building, as well as real-time meteorological data from the outside. Based on this, a directed energy transfer graph characterizing the electromechanical supply and demand logic is constructed. This directed energy transfer graph refers to a network structure that clearly shows the sequential connection relationships and flow directions of energy input and output between various electromechanical devices such as chillers, pumps, and terminal air conditioners. Next, the system extracts the energy efficiency decay rate of electromechanical nodes and the cross-node heat transfer delay from this structure. The energy efficiency decay rate reflects the decrease in actual energy conversion efficiency caused by the increase in the service life of aging equipment, while the heat transfer delay reflects the objective physical time lag phenomenon that exists when hot and cold water are transmitted in complex pipe networks. Finally, these parameters reflecting the actual aging status of equipment and the hydraulic hysteresis characteristics of the pipe network are integrated to generate a dynamic baseline energy consumption feature set.

[0022] Step S2 is primarily used for dynamic virtual trial and error and operational simulation of the renovation plan within the digital twin space. The system maps pre-set renovation plan parameters, such as replacing with a new high-efficiency host or adjusting control strategies, onto the previously established directed energy transfer graph to reconstruct the network structure, thereby generating a digital twin simulation map. Based on this map, the system performs multi-device coupled load extrapolation. Multi-device coupled load extrapolation refers to fully considering how changes to a local device in the system trigger chain reactions of energy consumption fluctuations in other upstream and downstream devices, and calculating the overall load change process accordingly. Through this cascaded extrapolation, the system can accurately extract the predicted energy efficiency improvement sequence and load time-series offset sequence for each electromechanical node. The load time-series offset sequence records in detail the shift trajectory of peak building electricity consumption on the time axis caused by the introduction of energy storage devices or optimized scheduling strategies. The purpose of this step is to break through the limitations of previous isolated energy-saving calculations for single devices, enabling early detection of potential hidden thermoelectric conflicts during multi-system joint renovations, and accurately predicting the actual energy-saving gain and load regulation capacity of the system after renovation.

[0023] Step S3 is primarily used to quantify the economic resilience and demand-side response potential of the retrofit scheme under complex external environments. The system first obtains an economic disturbance matrix containing fluctuations in carbon emission trading prices and time-of-use electricity prices, as well as a climate disturbance matrix containing the probability of extreme high or low temperatures. Then, it combines various feature sequences representing the pre- and post-retrofit states with the two time-varying disturbance matrices and substitutes them into a Markov chain Monte Carlo model for sampling evolution. The Markov chain Monte Carlo model is a statistical algorithm that uses multiple random sampling to simulate various complex and uncertain future scenarios and solve for probability distributions. Through extensive random walk evolution, the system analyzes the cost-effectiveness probability distribution set of the scheme throughout its entire lifecycle and the system's flexible response boundary. The system's flexible response boundary defines the maximum electrical load regulation extreme value that the building can withstand without compromising basic indoor thermal comfort.

[0024] Step S4 is primarily used to generate the optimal engineering execution decision that balances multiple business indicators while meeting the risk threshold. The system extracts the cost-effectiveness probability distribution set and system flexible response boundary from the output of step S3 to construct a multidimensional carbon economic evaluation feature set containing expected investment return indicators and grid interaction resilience indicators. Subsequently, the system performs non-dominated ranking optimization on this multidimensional carbon economic evaluation feature set. This non-dominated ranking optimization refers to the algorithm's ability to select a set of equilibrium solutions with the best overall benefits and no absolute disadvantages among multiple mutually constraining business objectives, such as high upfront investment and long-term carbon reduction benefits. Finally, the system selects the optimal solution that meets the project's pre-set business risk threshold and outputs the optimal retrofit decision map. The purpose of this step is to directly transform the extremely complex physical deductions and economic sampling data from the initial stages into an execution path that engineering installation companies or project investors can directly refer to, helping decision-makers find the safest investment balance between energy conservation and emission reduction requirements and project profitability goals.

[0025] Specifically, the process of acquiring the measured operation sequence and meteorological sequence of the building to be renovated, and constructing a directed graph representing the electromechanical supply and demand logic, is as follows: Collect multi-dimensional operating status parameters of the chiller units, water pumps, and air conditioning terminals at the bottom layer of the IoT system of the building to be renovated; simultaneously collect outdoor temperature and humidity parameters from the external meteorological monitoring terminal; perform multi-source time series timestamp alignment to generate the measured operation sequence and meteorological sequence of the building to be renovated; analyze the electromechanical-hydraulic connection topology logic of the chiller units, water pumps, and air conditioning terminals within the measured operation sequence of the building to be renovated; establish nodes and weighted edges that map the electromechanical-hydraulic connection topology logic; integrate nodes, weighted edges, the measured operation sequence of the building to be renovated, and the meteorological sequence to perform graph mapping structure assembly to generate a directed graph representing the electromechanical supply and demand logic.

[0026] In this implementation plan, the system first collects operating status parameters of the chiller unit and various terminal devices, as well as environmental parameters such as outdoor temperature and humidity. Due to the physical differences in the sampling frequencies of different sensors, the system directly resamples and timestamps various heterogeneous time-series data based on a unified clock reference sequence, thereby transforming the discrete raw data into a uniform measured operating sequence and meteorological sequence of the building to be modified. Subsequently, the system analyzes the fluid piping and electrical connection topology between devices, setting the chiller unit as the source node, the water pump as the transmission node, and the air conditioning terminal as the sink node, and establishes weighted edges connecting adjacent nodes. To accurately quantify the physical coupling strength of the pipe network connection, the system calculates the energy transmission coupling weight of each weighted edge. The calculation logic for the energy transmission coupling weight is to extract the source node output power sequence and the sink node received power sequence within a time window, calculate the absolute value of the difference between the two, and perform exponential decay mapping by combining the inherent damping coefficient of the pipe network. The specific formula is as follows: , in the formula This represents the energy transfer coupling weight from node m to node n; Indicates the inherent damping coefficient of the pipeline network; Represents an exponential function with the natural constant as its base; Indicates the spatiotemporal alignment adjustment factor; This represents the output thermal power eigenvector of node m; This represents the characteristic vector of the received heat power of node n; the spatiotemporal alignment adjustment factor is determined by calculating the reciprocal of the product of the square of the distance between nodes and the fluid velocity in the pipe network under historical benchmark tests. The above steps transform the complex building electromechanical system into a digital map, clearly characterizing the actual flow direction and dynamic loss patterns of energy between various equipment nodes, providing a high-fidelity physical network foundation for subsequent digital twin simulations.

[0027] Specifically, the process of extracting the energy efficiency decay rate and heat transfer delay of electromechanical nodes in the directed energy transfer graph and constructing a dynamic baseline energy consumption feature set is as follows: Track the deviation between the operating state of the energy efficiency benchmark data on the nameplates of the nodes in the directed energy transfer graph and the measured load response parameters, and analyze the energy efficiency decay rate of the electromechanical nodes; perform a time-series traversal of the cross-correlation function on the input thermal peak data and output thermal valley data of adjacent nodes in the directed energy transfer graph, locate the waveforms and align the corresponding time delay intervals, and extract the heat transfer delay across nodes; splice the energy efficiency decay rate, heat transfer delay, and measured operating sequence of the building to be renovated, and perform tensor fusion to generate a dynamic baseline energy consumption feature set.

[0028] In this implementation plan, the system first tracks the factory energy efficiency benchmark data marked on the nameplates of each electromechanical node and compares it with the measured energy conversion parameters under the current actual operating load. The deviation between the two operating states is measured to analyze the energy efficiency degradation rate caused by equipment aging. Subsequently, considering the hysteresis characteristics of the fluid pipeline network, the system extracts the input thermal peak data and output thermal trough data between adjacent nodes and performs a cross-correlation function time-series traversal. The purpose of the cross-correlation function time-series traversal is to compare the waveform similarity of two time-series signals through a sliding time window, thereby determining the exact physical time required for heat to be transferred from upstream to downstream. The specific calculation formula is as follows: , in the formula This indicates that nodes x and y have a time delay of... The cross-correlation coefficient at time; Z represents the total sample length of the sliding time window; t represents the current discrete time step; This represents the input thermodynamic feature sequence value of node x at time step t; Indicates node y at time step The system outputs thermal characteristic sequence values; it iterates through the set optimization interval to obtain the time delay corresponding to the maximum value of the cross-correlation coefficient, and compares it with a preset hysteresis significance threshold. If it is greater than the hysteresis significance threshold, it is extracted as the cross-node heat and cold transfer delay; otherwise, it is considered zero delay. The hysteresis significance threshold is determined by extracting the average of the minimum pipeline filling response time recorded in multiple cold start tests. Finally, the system tensors and fuses indicators reflecting the performance degradation of equipment and pipelines, such as energy efficiency decay rate and heat and cold transfer delay, with the measured building operation sequence along the time dimension to generate a dynamic baseline energy consumption feature set. This process accurately quantifies the current health status of the equipment, truly restores the dynamic delay effect of pipeline heat transfer, and allows the evaluation model to completely break free from the limitations of ideal operating conditions, greatly improving the physical accuracy of building energy consumption baseline prediction.

[0029] Specifically, the process of mapping the preset modification scheme parameters into the directed energy transfer graph for reconstruction and generating a digital twin simulation graph is as follows: Analyze the new host performance characteristics, pipeline resistance variation characteristics, and control scheduling variation strategies embedded in the preset modification scheme parameters; unload high-energy-consumption aging nodes from the directed energy transfer graph, implant new nodes corresponding to the new host performance characteristics, and refresh the weighted edge connectivity resistance matrix of the directed energy transfer graph according to the pipeline resistance variation characteristics; transform the control scheduling variation strategy into node collaborative scheduling constraints, reconstruct the topological mapping relationship of the directed energy transfer graph, and generate the digital twin simulation graph.

[0030] In this implementation plan, the analysis of the embedded performance characteristics of the new host, the characteristics of pipeline resistance variation, and the control scheduling variation strategy within the preset modification scheme parameters aims to transform the physical plan on the design drawings into identifiable underlying constraint data in the digital twin space. The system addresses and removes nodes representing high-energy-consuming, outdated equipment within the directed energy transfer graph, simultaneously embedding virtual new nodes corresponding to the performance characteristics of the new host. Since equipment replacement inevitably causes changes in hydraulic distribution, the system recalculates and refreshes the weighted edge connectivity resistance matrix based on the characteristics of pipeline resistance variation. The specific calculation logic is to subtract the resistance compensation value brought by frequency conversion regulation from the product of the pipeline material roughness and the design parameters of the two end nodes. The calculation formula is as follows: , in the formula This represents the updated connectivity resistance characteristic value between node k and node l; Indicates the roughness coefficient of the pipeline material; This represents the design head parameter of the new host node k; This represents the required water pressure parameter for receiving node l; This represents the variable frequency control compensation weighting coefficient; Represents the baseline flow velocity at node k; Represents a node The reference flow rate; This represents the function that takes the maximum value of the two options; the method for determining the frequency conversion regulation compensation weight coefficient is to extract the square of the ratio of the frequency converter output frequency to the rated frequency under the historical best operating condition for the same pipe diameter. Subsequently, the system converts the new control scheduling change strategy into the collaborative scheduling boundary conditions for the state transition of the constraint graph nodes, completing a comprehensive reconstruction of the topology mapping relationship, and finally generating an accurate digital twin simulation graph of the mapping and renovation design. The purpose of this step is to ensure that the virtual simulation network and the actual renovation scheme remain absolutely consistent at the physical constraint level, thereby providing an accurate network structure framework for subsequent system-level energy consumption calculations.

[0031] Please see Figure 2 Specifically, the process of performing multi-device coupled load extrapolation based on digital twin simulation maps and extracting the energy efficiency improvement prediction sequence and load time-series offset sequence of electromechanical nodes is as follows: Apply historical full-condition heat load boundary conditions to the digital twin simulation map, trigger the full-network heat and power coordinated cascade response along the weighted edge direction, and perform multi-device coupled load extrapolation; collect the dynamic energy efficiency parameters of the nodes generated by the multi-device coupled load extrapolation, map and compare them with the corresponding nodes in the dynamic baseline energy consumption feature set to extract the gain amplitude features, and establish the energy efficiency improvement prediction sequence of electromechanical nodes; track the timing span characteristics of the cold storage and release of nodes triggered by the control and scheduling change strategy, analyze the trajectory of the electricity load peak on the time axis, and construct the load time-series offset sequence.

[0032] In this implementation plan, the purpose of performing multi-device coupled load simulation based on the digital twin simulation map is to predict the actual energy-saving level and grid peak-shaving potential of the renovation scheme in a real complex physical environment. The system applies heat load boundary conditions covering the entire historical operating conditions of the building to the reconstructed digital twin simulation map, driving the conduction of cooling and heating along the weighted edge direction, triggering a cascaded response of the entire network's thermal and electrical systems from the main unit to the terminal. During the cascaded simulation, the system frequently collects the dynamically evolved energy efficiency parameters of each node and maps and compares them with the corresponding baseline nodes to extract gain amplitude characteristics. The specific calculation process involves multiplying the virtual operating energy efficiency parameters by the coupling correction coefficient, scaling the parameters to the negative first power of the baseline parameters, and then subtracting the error compensation. The formula is as follows: , in the formula This represents the gain amplitude characteristic value of the electromechanical node u at the operating step size h; This represents the virtual operating energy efficiency parameter of node u generated by multi-device coupling simulation; This represents the original energy efficiency parameters corresponding to the dynamic baseline energy consumption feature set; This represents the coupling and coordination correction coefficient for node u; This represents the compensation value for the truncation error in the model simulation; the method for determining the coupling and collaborative correction coefficient is to extract the logarithm of the product of the rated power of all series-connected electromechanical equipment on the hydraulic branch where node u is located. Based on the above calculation system, a predictive sequence of energy efficiency improvement of electromechanical nodes over time is established. At the same time, the system tracks the time span of the execution nodes of control strategies such as cold storage and release, and analyzes the translation trajectory of the load peak on the time axis to construct a load time-series offset sequence. This step, by introducing a full-condition boundary and cascade response mechanism, enables energy efficiency assessment to overcome the limitations of isolated static calculations of a single machine, and can accurately capture the nonlinear energy-saving fluctuations caused by the mutual influence between equipment.

[0033] Specifically, the process of obtaining the economic disturbance matrix containing carbon electricity price fluctuations and the climate disturbance matrix containing extreme weather probabilities is as follows: Collect historical settlement price time-series data from the carbon emission trading market and time-of-use electricity price transition time-series data from the regional power grid; analyze the frequency domain characteristics and confidence interval boundaries of price fluctuations to generate the economic disturbance matrix; obtain historical extreme high and low temperature frequency samples corresponding to the geographical coordinates of the building to be renovated; fit the probability density function of extreme temperature and humidity meteorological elements; extract discrete meteorological mutation vectors exceeding the baseline climate conditions to generate the climate disturbance matrix; align the time step of price evolution within the economic disturbance matrix with the execution timestamp of mutation occurrence time step within the climate disturbance matrix to output the economic disturbance matrix and the climate disturbance matrix.

[0034] In this implementation plan, the acquisition of the economic disturbance matrix and the climate disturbance matrix is ​​mainly used to establish the dynamic external risk boundaries faced by the energy efficiency assessment model over a long period. Here, the economic disturbance matrix and the climate disturbance matrix represent the nonlinear impact of future carbon trading market fluctuations and frequent extreme weather events on building operating costs and load demand. Specifically, the system collects historical settlement price time-series data from the carbon emission trading market and step transition time-series data from the regional power grid's time-of-use electricity prices, analyzing the frequency domain characteristics and confidence interval boundaries of price fluctuations. To accurately quantify the impact of price fluctuations, the system constructs an economic disturbance matrix. The calculation logic for the elements in the matrix is ​​to multiply the base fluctuation weight coefficient by the benchmark price and perform an exponential mapping based on the frequency domain amplification factor and transition amplitude. The specific formula is as follows: , in the formula This represents the economic disturbance coefficient for disturbance dimension a at time step b; This represents the basic fluctuation weighting coefficient for dimension a; This represents the carbon trading settlement benchmark price at time step b; Represents an exponential function with the natural constant as its base; This represents the frequency domain magnification factor of dimension a; This represents the time-of-use electricity price jump at time step b; the basic fluctuation weighting coefficient is determined by extracting the absolute value of the difference between the maximum peak and minimum trough in the corresponding dimension over five consecutive years. Subsequently, the system obtains historical extreme high and low temperature frequency samples corresponding to the geographical coordinates of the building to be renovated, fits the probability density function of extreme temperature and humidity meteorological elements, and extracts discrete meteorological abrupt change vectors that exceed conventional benchmark climate conditions, thereby generating a climate perturbation matrix. Finally, the system strictly aligns the time steps of both. This step effectively solves the defect of traditional assessments using static electricity prices and constant meteorological data, which leads to distorted calculation results, and injects the system with real external market and natural environmental evolution patterns.

[0035] Please see Figure 3 Specifically, the process of substituting the dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time-series offset sequence, economic disturbance matrix, and climate disturbance matrix into the Markov chain Monte Carlo model for sampling evolution, and extracting the cost-effectiveness probability distribution set and the system flexible response boundary is as follows: The dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, and load time-series offset sequence are mapped to the initial state space of the Markov chain Monte Carlo model; the economic disturbance matrix and climate disturbance matrix are transformed into random walk state transition probability parameters of the Markov chain Monte Carlo model; the Markov chain Monte Carlo model is triggered to perform a cross-lifecycle time-step random sampling walk; the initial state space and state transition probability parameters are fused to perform dimensionality reduction of carbon assets and energy consumption input features; the net present value evolution trajectory under the clustering convergence state is clustered to generate the cost-effectiveness probability distribution set; the extreme values ​​of the load time-series offset response are tracked under the limit state containing discrete meteorological mutation vectors during the random sampling walk; the maximum load peak-shaving and valley-filling capacity of electromechanical network nodes under the condition of not violating thermal comfort constraints is analyzed; and the system flexible response boundary is extracted.

[0036] In this implementation plan, various feature sequences and perturbation matrices are substituted into a Markov chain Monte Carlo model for sampling evolution, primarily to explore the optimal investment return probability and grid interaction potential of the transformation scheme under massive uncertainty scenarios. The Markov chain Monte Carlo model here is a complex statistical algorithm that performs multidimensional random walks across the entire state space based on the current initial state and state transition probability matrix. The system first maps the previously generated dynamic baseline energy consumption feature set and energy efficiency improvement prediction sequence to the initial state space of the model, while simultaneously converting the economic and climate perturbation matrices into state transition probability parameters for the model's random walks at each time step. Subsequently, the model is triggered to perform multiple random sampling walks throughout the simulated entire lifecycle, reducing the dimensionality and clustering the carbon asset and energy consumption input characteristics to generate a cost-effectiveness probability distribution set for the net present value evolution trajectory. During this process, the system simultaneously tracks the load response extreme values ​​under the extreme state of encountering discrete meteorological mutation vectors and calculates the system's flexible response boundary. The specific calculation logic is to calculate the sum of the rated power of each electromechanical node after experiencing thermal comfort attenuation penalties and subtract the basic maintenance energy consumption, as shown in the formula below. , in the formula This represents the system's maximum load peak shaving and valley filling capacity under a sudden weather change condition (c). The parameter representing the state transition probability of a random walk in state c; This represents the summation function; D represents the total number of nodes in the electromechanical pipeline network. This indicates the rated operating power of node d in the electromechanical pipeline network; This represents the thermal comfort sensitivity attenuation factor at node d; This represents the evolution value of the outdoor extreme temperature under a sudden meteorological state c; This represents the indoor thermal comfort temperature limit for the area served by node d; This represents the system's basic maintenance energy consumption compensation threshold; where the system's basic maintenance energy consumption compensation threshold... The method for determining this is to extract the average energy consumption across the entire building when it is unattended at night and all equipment is at its lowest standby frequency. This step completely changes the traditional, single, and deterministic energy-saving accounting model, expanding the assessment dimensions to probabilistic risk management and flexible electricity consumption boundary exploration throughout the entire life cycle, greatly enhancing the commercial reliability and engineering feasibility of the final renovation decision in complex environments.

[0037] Please see Figure 4Specifically, the process of constructing a multidimensional carbon economy evaluation feature set based on the cost-effectiveness probability distribution set and the system flexible response boundary, and performing non-dominated ranking optimization on the multidimensional carbon economy evaluation feature set to output the optimal transformation decision map that meets the preset business risk threshold is as follows: The cost-effectiveness probability distribution set is dimensionality-reduced and mapped to an expected investment return feature vector; the system flexible response boundary is transformed into a grid interaction elasticity feature vector; the expected investment return feature vector and the grid interaction elasticity feature vector are fused to generate the multidimensional carbon economy evaluation feature set; a preset business risk threshold is set as the optimization constraint boundary; the multidimensional carbon economy evaluation feature set is substituted into the non-dominated ranking algorithm space to perform Pareto front dominance hierarchical filtering, extracting the optimal Pareto front solution set that fits the optimization constraint boundary; the corresponding transformation equipment combinations and control logic within the optimal Pareto front solution set are mapped to a visualized node relationship network to generate the optimal transformation decision map.

[0038] In this implementation scheme, a multidimensional carbon economic evaluation feature set is constructed based on the cost-effectiveness probability distribution set and the system flexible response boundary, and then optimized and output. This is primarily used to screen the optimal engineering execution path from a vast number of potential retrofit schemes, balancing high economic returns with high risk resistance. The system first reduces the dimensionality of the large-scale cost-effectiveness probability distribution set generated in the preceding steps, mapping it to a single-dimensional expected investment return feature vector. Simultaneously, the system flexible response boundary is transformed into a grid interaction elasticity feature vector characterizing the potential of buildings to participate in the grid demand-side response. Subsequently, the system merges these two types of feature vectors to generate a multidimensional carbon economic evaluation feature set, setting a preset commercial risk threshold as a strict optimization constraint boundary. The system substitutes the feature set into the non-dominated ranking algorithm space to perform Pareto front-dominated hierarchical filtering. Here, non-dominated ranking refers to a statistical algorithm that performs balanced evaluation and ranking among multiple mutually constraining objectives, including investment returns and grid elasticity. The Pareto front represents the theoretically optimal solution set under current constraints where one objective cannot be further optimized without sacrificing another. The specific hierarchical filtering logic involves subtracting a risk exceeding the boundary penalty term from the product of the expected feature vector and the elasticity feature vector. The specific calculation formula is as follows: , in the formula This represents the overall superiority rating score of the modified solution set p; This represents the economic return weighting coefficient; This represents the magnitude of the eigenvector representing the expected return on investment for the solution set p after renovation. This represents the magnitude of the power grid interaction elasticity eigenvector of the modified solution set p; This indicates a penalty factor for exceeding risk limits; Represents an exponential function with the natural constant as its base; This represents the assessment of business risk index for the modified solution set p; This represents the preset business risk threshold; where the preset business risk threshold... The method for determining the optimal Pareto front is to obtain the average risk tolerance limit value of investment loss boundaries in similar building renovation projects in the same region over the past five years. Based on this comprehensive superiority rating score, the system extracts the optimal Pareto front solution set that perfectly fits the constraint boundary. Finally, the specific equipment model combinations and operational control logic within the solution set are mapped to a visual node relationship network, outputting the final optimal renovation decision map. This step effectively transforms the complex physical simulation and probabilistic extrapolation data into intuitive decision-making data that engineers can readily understand. It effectively avoids the blind spot of solely pursuing high energy efficiency while ignoring market fluctuation risks, ensuring the economic robustness and technical feasibility of green building renovation solutions in actual commercial operation.

[0039] Example 2; please refer to Figure 5 A digital twin energy efficiency assessment system for green building renovation, used to execute the digital twin energy efficiency assessment method for green building renovation described in Example 1, includes: a baseline extraction module, used to acquire the measured operation sequence and meteorological sequence of the building to be renovated, and construct a directed energy transfer graph representing the electromechanical supply and demand logic; extract the energy efficiency decay rate and heat transfer delay of the electromechanical nodes in the directed energy transfer graph, and construct a dynamic baseline energy consumption feature set; a twin simulation module, used to map the preset renovation scheme parameters into the directed energy transfer graph for reconstruction, and generate a digital twin simulation map; perform multi-device coupled load simulation based on the digital twin simulation map, and extract the energy efficiency improvement prediction of the electromechanical nodes. The system includes a measurement sequence and a load time-series offset sequence; a sampling evolution module, used to obtain an economic disturbance matrix containing carbon electricity price fluctuations and a climate disturbance matrix containing extreme weather probabilities; a dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time-series offset sequence, economic disturbance matrix, and climate disturbance matrix are substituted into a Markov chain Monte Carlo model for sampling evolution, and the cost-effectiveness probability distribution set and system flexible response boundary are extracted analytically; and a decision optimization module, used to construct a multidimensional carbon economic evaluation feature set based on the cost-effectiveness probability distribution set and system flexible response boundary, perform non-dominated ranking optimization on the multidimensional carbon economic evaluation feature set, and output the optimal transformation decision map that meets the preset business risk threshold.

[0040] In this implementation plan, the baseline extraction module is primarily responsible for the initial data acquisition and processing tasks to construct a realistic physical foundation for building energy consumption. This module automatically accesses and integrates multi-dimensional sensor data from the building's underlying electromechanical equipment under actual operating conditions, as well as real-time environmental data from external weather stations, abandoning the previous crude method of relying solely on ideal parameters from design drawings. It can accurately extract the actual degradation of equipment over time and the actual physical hysteresis of hydraulic transmission in the pipeline network, thus providing the entire assessment system with a dynamic baseline that is completely faithful to the building's current health status. This ensures that all subsequent retrofit gain calculations are based on a truly reliable engineering starting point.

[0041] The twin simulation module is the core execution component of the entire evaluation system, used for virtual trial and error and scheme pre-simulation. This module is responsible for accurately projecting the various electromechanical modification parameters pre-determined by engineers onto the baseline network, quickly constructing a virtual model consistent with the future physical state after modification. Its greatest advantage lies in breaking the traditional limitation of calculating energy efficiency in isolation for individual electromechanical devices. It can simulate in advance in digital space the chain reaction of thermal and electrical effects on the entire system triggered by the connection of new equipment. This helps the implementation team understand the actual energy efficiency improvement potential and power load transfer capacity under multi-device linkage conditions before the project actually starts, avoiding system-level mismatches during actual on-site modifications.

[0042] The sampling evolution module serves as a risk testing center for the renovation project in the face of future complex environmental changes. This module proactively incorporates multiple external environmental and market disturbances that may be encountered in the long term, such as fluctuations in carbon emission market trading prices, regional time-of-use electricity price adjustments, and frequent extreme high and low temperature weather. By performing massive random sampling calculations, this module can transform the originally single, fixed static investment return prediction into a financial risk assessment distribution that covers the probability of various complex scenarios. It also calculates the electricity consumption regulation limits of the building's pipeline network to maintain basic thermal comfort under severe weather conditions, providing a quantitative assessment of the system's economic security throughout its entire lifecycle.

[0043] The decision optimization module serves as the intelligent adjudication hub responsible for outputting the final project execution plan. Faced with numerous conflicting economic return and risk resilience indicators generated from preliminary simulations, this module rigorously filters and prioritizes various parameters based on the investor's or contractor's pre-set commercial risk tolerance threshold. It simplifies and visualizes the results of complex high-dimensional algorithms, directly selecting the equipment combinations and control strategies that best meet the current project's dual requirements of profitability and safety. This is output in the form of an intuitive network diagram, providing scientific and reliable optimal decision-making guidance for construction and installation companies' procurement negotiations and on-site project construction.

[0044] In summary, this application has at least the following effects: A digital twin energy efficiency assessment method and system for green building renovation first constructs a directed energy transfer graph by collecting real IoT operation sequences and meteorological sequences. This accurately quantifies the actual energy efficiency degradation rate of aging electromechanical equipment and the delay in heat and cold transfer in the pipeline network, overcoming the shortcomings of traditional assessment methods that rely on static design parameters, leading to distorted energy consumption benchmarks. Then, the renovation scheme parameters are mapped onto the digital twin simulation graph to perform multi-device coupled load extrapolation, effectively predicting the overall energy efficiency improvement potential of the pipeline network and the trajectory of electricity load shift after the new equipment is connected. This avoids the energy-saving expectation errors caused by isolated calculations of individual electromechanical equipment. Finally, a carbon trading market is introduced. Real-world environmental disturbances such as fluctuating electricity prices and the probability of extreme weather events are addressed by using a Markov chain Monte Carlo model for multidimensional random sampling evolution across the lifecycle. This quantifies and analyzes the cost-effectiveness probability distribution of the retrofit scheme and the grid's flexible response boundary, effectively solving the economic pain point that relying on static financial accounting with a constant unit price can easily lead to real project investment losses. Finally, non-dominated ranking optimization is performed on the multidimensional carbon economic evaluation feature set to directly output the optimal retrofit decision map that meets the project's commercial risk tolerance threshold. This provides a scientific implementation path for green building electromechanical retrofit projects that balances high energy-saving and emission-reduction returns with high risk resistance.

[0045] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A digital twin energy efficiency assessment method for green building renovation, characterized in that, Includes the following steps: S1. Obtain the measured operation sequence and meteorological sequence of the building to be modified, and construct a directed energy transfer graph representing the electromechanical supply and demand logic; Extract the energy efficiency decay rate and heat transfer delay of electromechanical nodes in the directed energy transfer graph to construct a dynamic baseline energy consumption feature set; S2. Map the preset modification scheme parameters into the directed energy transfer graph for reconstruction, and generate a digital twin simulation map; Based on the digital twin simulation map, multi-device coupled load extrapolation is performed, and the energy efficiency improvement prediction sequence and load time series offset sequence of electromechanical nodes are extracted; S3. Obtain the economic disturbance matrix containing carbon electricity price fluctuations and the climate disturbance matrix containing extreme weather probability. Substitute the dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time series offset sequence, economic disturbance matrix and climate disturbance matrix into the Markov chain Monte Carlo model for sampling evolution, and analyze and extract the cost-effectiveness probability distribution set and system flexible response boundary. S4. Construct a multidimensional carbon economy evaluation feature set based on the cost-effectiveness ratio probability distribution set and the system flexible response boundary. Perform non-dominated ranking optimization on the multidimensional carbon economy evaluation feature set and output the optimal transformation decision map that meets the preset business risk threshold.

2. The method for evaluating the energy efficiency of a digital twin for green building renovation according to claim 1, characterized in that: The specific process of obtaining the measured operation sequence and meteorological sequence of the building to be renovated, and constructing a directed graph representing the electromechanical supply and demand logic is as follows: Collect multi-dimensional operating status parameters of the chiller, water pump and air conditioning terminal of the building IoT system to be renovated, and simultaneously collect outdoor temperature and humidity parameters of the external meteorological monitoring terminal. Perform multi-source time series timestamp alignment to generate the measured operating sequence and meteorological sequence of the building to be renovated. The topological association logic of the electromechanical and hydraulic connections of chiller units, water pumps and air conditioning terminals in the measured operation sequence of the building to be renovated is analyzed, and the nodes and weighted edges that map the topological association logic of the electromechanical and hydraulic connections are established. By integrating nodes, weighted edges, measured operation sequences of buildings to be modified, and meteorological sequences into a mapping structure, a directed energy transfer graph representing the electromechanical supply and demand logic is generated.

3. The method for evaluating the energy efficiency of a digital twin for green building renovation according to claim 1, characterized in that: The specific process of extracting the energy efficiency decay rate and heat transfer delay of electromechanical nodes in the directed energy transfer graph and constructing a dynamic baseline energy consumption feature set is as follows: The deviation between the energy efficiency benchmark data on the nameplate of the energy transfer directed graph node and the measured load response parameters is tracked to analyze the energy efficiency degradation rate of the electromechanical node. Perform a time-series cross-correlation function on the input thermal peak data and output thermal valley data of adjacent nodes in the directed graph of energy transfer, locate the waveform alignment corresponding time delay interval, and extract the cross-node heat transfer delay. The energy efficiency decay rate, heat transfer delay, and measured operating sequence of the building to be renovated are spliced ​​together and tensor fusion is performed to generate a dynamic baseline energy consumption feature set.

4. The method for evaluating the energy efficiency of a digital twin for green building renovation according to claim 1, characterized in that: The specific process of mapping the preset modification scheme parameters into the directed energy transfer graph for reconstruction and generating a digital twin simulation map is as follows: The performance characteristics of the new host, the characteristics of pipeline resistance variation, and the control and scheduling variation strategies embedded in the parameters of the preset modification scheme are analyzed. Unload high-energy-consuming aging nodes in the directed energy transfer graph, implant new nodes corresponding to the performance characteristics of new host, and refresh the weighted edge connectivity resistance matrix of the directed energy transfer graph according to the characteristics of pipeline resistance variation. The control and scheduling change strategy is transformed into node collaborative scheduling constraints, and the energy transfer directed graph topology mapping relationship is reconstructed to generate a digital twin simulation map.

5. The method for evaluating the energy efficiency of a digital twin for green building renovation according to claim 4, characterized in that: The specific process of performing multi-device coupled load extrapolation based on digital twin simulation maps and extracting the energy efficiency improvement prediction sequence and load time series offset sequence of electromechanical nodes is as follows: Apply historical full-condition heat load boundary conditions to the digital twin simulation map, trigger the whole network heat and power coordinated cascade response along the weighted edge direction, and perform multi-equipment coupled load simulation; Collect dynamic energy efficiency parameters of nodes generated by multi-device coupled load simulation, map and compare them with corresponding nodes in the dynamic baseline energy consumption feature set to extract gain amplitude features, and establish a prediction sequence for energy efficiency improvement of electromechanical nodes. By tracking the timing characteristics of the cold storage and release time span of the node triggered by the change strategy of control and scheduling, the trajectory of the load peak on the time axis is analyzed, and a load time offset sequence is constructed.

6. The method for evaluating the energy efficiency of a digital twin for green building renovation according to claim 1, characterized in that: The specific process for obtaining the economic disturbance matrix including carbon electricity price fluctuations and the climate disturbance matrix including extreme weather probabilities is as follows: Collect historical settlement price time series data of the carbon emission trading market and time series data of the time-of-use electricity price tier transition of the regional power grid, analyze the frequency domain characteristics of price fluctuations and confidence interval boundaries, and generate an economic disturbance matrix; Obtain historical extreme high and low temperature frequency samples corresponding to the geographical coordinates of the building to be renovated, fit the probability density function of extreme temperature and humidity meteorological elements, extract discrete meteorological abrupt change vectors that exceed the baseline climate conditions, and generate a climate perturbation matrix. Align the time step of price evolution within the economic disturbance matrix with the execution timestamp of abrupt change occurrence within the climate disturbance matrix, and output the economic disturbance matrix and the climate disturbance matrix.

7. The method for evaluating the energy efficiency of a digital twin for green building renovation according to claim 1, characterized in that: The specific process of substituting the dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time series offset sequence, economic disturbance matrix, and climate disturbance matrix into the Markov chain Monte Carlo model for sampling evolution, and then analyzing and extracting the cost-effectiveness probability distribution set and the system flexible response boundary is as follows: The dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence and load time series offset sequence are mapped to the initial state space of the Markov chain Monte Carlo model, and the economic disturbance matrix and climate disturbance matrix are transformed into random walk state transition probability parameters of the Markov chain Monte Carlo model. The Markov chain Monte Carlo model is triggered to perform a random sampling walk across the life cycle time step, and the initial state space and state transition probability parameters are integrated to perform dimensionality reduction of carbon asset and energy consumption input characteristics, cluster the net present value evolution trajectory in the convergent state, and generate a cost-effectiveness probability distribution set. By tracking the extreme values ​​of the load time-series offset response under the limit state containing discrete meteorological change vectors during random sampling walks, the maximum load peak-shaving and valley-filling capacity of electromechanical network nodes without violating thermal comfort constraints is analyzed, and the flexible response boundary of the system is extracted.

8. The method for evaluating the energy efficiency of a digital twin for green building renovation according to claim 1, characterized in that: The specific process of constructing a multidimensional carbon economy evaluation feature set based on the cost-effectiveness ratio probability distribution set and the system flexible response boundary, performing non-dominated ranking optimization on the multidimensional carbon economy evaluation feature set, and outputting the optimal transformation decision map that meets the preset business risk threshold is as follows: The cost-effectiveness probability distribution set is reduced in dimension and mapped to the expected return on investment feature vector, and the system flexible response boundary is transformed into the power grid interactive elasticity feature vector; By integrating the expected investment return feature vector and the power grid interaction elasticity feature vector, a multidimensional carbon economy evaluation feature set is generated. A preset business risk threshold is set as the optimization constraint boundary. The multidimensional carbon economy evaluation feature set is substituted into the non-dominated sorting algorithm space to perform Pareto front dominance hierarchical filtering and extract the optimal Pareto front solution set that fits the optimization constraint boundary. The optimal Pareto front solution set is mapped to the corresponding combination of modified equipment and control logic as a visual node relationship network, generating an optimal modification decision graph.

9. A digital twin energy efficiency assessment system for green building renovation, used to execute the digital twin energy efficiency assessment method for green building renovation as described in any one of claims 1-8, characterized in that, include: The baseline extraction module is used to obtain the measured operation sequence and meteorological sequence of the building to be modified, and to construct a directed graph of energy transfer that represents the electromechanical supply and demand logic; Extract the energy efficiency decay rate and heat transfer delay of electromechanical nodes in the directed energy transfer graph to construct a dynamic baseline energy consumption feature set; The twin simulation module is used to map the parameters of the preset modification scheme into the directed energy transfer graph for reconstruction, and generate a digital twin simulation map. Based on the digital twin simulation map, multi-device coupled load extrapolation is performed, and the energy efficiency improvement prediction sequence and load time series offset sequence of electromechanical nodes are extracted; The sampling evolution module is used to obtain the economic disturbance matrix containing carbon electricity price fluctuations and the climate disturbance matrix containing extreme weather probabilities. It substitutes the dynamic baseline energy consumption feature set, energy efficiency improvement prediction sequence, load time series offset sequence, economic disturbance matrix and climate disturbance matrix into the Markov chain Monte Carlo model for sampling evolution, and analyzes and extracts the cost-effectiveness probability distribution set and the system flexible response boundary. The decision optimization module is used to construct a multidimensional carbon economy evaluation feature set based on the cost-effectiveness probability distribution set and the system flexible response boundary. It performs non-dominated ranking optimization on the multidimensional carbon economy evaluation feature set and outputs the optimal transformation decision map that meets the preset business risk threshold.

Images