Energy-saving and environment-friendly green building design method

Abstract

The invention relates to the technical field of green building design, and discloses an energy-saving and environment-friendly green building design method. The method comprises the following steps: acquiring real-time monitoring data of a target building environment, and constructing an initial parameter set containing an energy consumption intensity gradient sequence and a material distribution curve; developing an energy consumption regulation and control model according to the energy consumption intensity gradient sequence, expressing the dynamic association between the energy consumption intensity and the building structure, forming an environmental protection regulation and control model based on a material distribution curve, and expressing the nonlinear relationship between material distribution and ecological indexes; in a design stage, two types of models are dynamically selected and an adjustment instruction is output according to structure deviation data and ecological index fluctuation data monitored in real time; and injecting an instruction into a design execution process to realize real-time correction of the strength gradient of the energy consumption device or the distribution curve of the material selection device. According to the method, multi-dimensional parameters are combined with a dynamic model, so that energy-saving and environment-friendly regulation and control of green building design better meet actual requirements.

Description

Technical Field

[0001] This invention relates to the field of green building design technology, specifically to an energy-saving and environmentally friendly green building design method. Background Technology

[0002] With the global energy crisis and ecological environmental problems becoming increasingly prominent, green building, as an important vehicle for reducing energy consumption and ecological impact, has seen its design methodology optimization become one of the core directions for industry development. Currently, in the green building design process, the control of energy conservation and environmental protection largely relies on preset fixed parameters or single-dimensional monitoring data, which is insufficient to comprehensively reflect the dynamic changes of the target building environment.

[0003] In terms of energy consumption control, existing technologies typically rely solely on basic building energy consumption data (such as electricity and water consumption per unit area) to establish control logic, without forming a parameter system that encompasses the energy intensity gradient sequence. This results in the control model being unable to accurately express the dynamic relationship between energy intensity and building structure (such as wall insulation layer thickness and window insulation performance). When environmental conditions such as temperature and light levels change, the control commands output by the model are prone to becoming disconnected from actual needs, leading to energy waste or a decrease in indoor comfort.

[0004] In terms of environmental control of materials, traditional methods mostly focus on the environmental indicators of the materials themselves (such as formaldehyde emission and recyclability), lacking in-depth analysis of the relationship between the distribution of materials in the overall building and ecological indicators (such as regional carbon emissions and the impact on surrounding vegetation). Most design schemes only meet basic environmental requirements through static material lists, without constructing models that can reflect the non-linear relationship between material distribution curves and ecological indicators. This makes it difficult to adapt material selection and layout to the ecological needs of the target environment. For example, using materials that are easily degraded by moisture in high-humidity areas, or overusing high-energy-consuming building materials in ecologically sensitive areas, exacerbates the environmental burden.

[0005] In the implementation phase of green building design, existing technologies generally lack dynamic response mechanisms for real-time data. When structural deviations (such as decreased wall sealing due to construction errors) or fluctuations in ecological indicators (such as short-term deterioration of surrounding air quality) are detected, adjustments to the design scheme often rely on manual judgment, failing to quickly select a suitable control model and output precise instructions. This lag not only affects the implementation of energy-saving and environmental protection effects but may also lead to repeated modifications to the design scheme, increasing time and economic costs. Furthermore, in existing methods, energy consumption control and environmental protection control often operate independently, failing to achieve synergy between the two. This makes it difficult to find a dynamic balance between energy consumption optimization and ecological protection, thus hindering the overall improvement of the energy-saving and environmental protection benefits of green buildings. Summary of the Invention

[0006] The purpose of this invention is to provide an energy-saving and environmentally friendly green building design method to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides an energy-saving and environmentally friendly green building design method, the method comprising:

[0008] Acquire real-time monitoring data of the target building environment, and construct an initial parameter set based on the real-time monitoring data. The initial parameter set includes an energy intensity gradient sequence and a material distribution curve.

[0009] Based on the energy intensity gradient sequence in the initial parameter set, an energy consumption control model that is coordinated with the target environment is developed. The energy consumption control model is used to express the dynamic relationship between energy intensity and building structure.

[0010] Based on the material distribution curves in the initial parameter set, an environmental protection regulation model that fits the target environment is formed. The environmental protection regulation model is used to express the nonlinear relationship between material distribution and ecological indicators.

[0011] During the green building design phase, based on real-time monitored structural deviation data and ecological indicator fluctuation data, an energy consumption control model or an environmental protection control model is dynamically selected, and adjustment instructions are output through the selected model.

[0012] Adjustment instructions are injected into the execution process of green building design to enable real-time correction of the intensity gradient of energy consumption devices or the distribution curve of material selection devices.

[0013] Preferably, the steps of acquiring real-time monitoring data and constructing an initial parameter set are further refined as follows:

[0014] Multiple environmental sensors are used to simultaneously collect meteorological parameters of the target environment during the planning period, including temperature change curves and humidity distribution data;

[0015] Based on the extreme regions in the temperature change curve, the target environment is divided into multiple sub-regions, and an initial energy consumption gradient range and material distribution reference value are set for each sub-region.

[0016] Retrieve historical energy consumption intensity compensation parameters and material distribution adjustment coefficients corresponding to each sub-region from the preset database;

[0017] By performing time-domain filtering on the historical energy consumption intensity compensation parameters, the optimized energy consumption intensity gradient sequence for each sub-region is obtained.

[0018] The material distribution adjustment coefficient is smoothed in the frequency domain to generate an optimized material distribution curve for each sub-region;

[0019] The optimized energy intensity gradient sequence and optimized material distribution curve of all sub-regions are integrated to form an initial parameter set.

[0020] Preferably, the steps for developing energy consumption control models and forming environmental protection control models include:

[0021] The optimized energy consumption intensity gradient sequence is input into the preset regression model. After multiple training iterations, the hidden layer weights are updated to generate the intensity compensation function in the energy consumption regulation model.

[0022] The optimized material distribution curve is input into the clustering model, and the cluster centers are optimized through the backpropagation algorithm to generate the distribution correction mapping table in the environmental protection regulation model.

[0023] After the regression model and clustering model converge, the feature vectors of the intensity compensation function and the distribution correction mapping table are extracted respectively.

[0024] The matching degree between the feature vector and real-time meteorological parameters is calculated to verify the applicability of the energy consumption control model and the environmental protection control model.

[0025] When the matching degree is lower than the set standard, the number of training iterations of the regression model and clustering model is readjusted until the feature vector meets the applicable conditions.

[0026] Preferably, the steps for dynamically selecting a model and outputting adjustment instructions are as follows:

[0027] Continuously track the changing trends of structural deviations and the oscillation range of ecological indicators during the design process;

[0028] When the structural deviation trend exceeds the first threshold and the ecological index oscillation is within a stable range, the intensity compensation function in the energy consumption regulation model is activated.

[0029] Based on the gradient adjustment rule in the intensity compensation function, an intensity correction command for the energy consumption device is generated.

[0030] When the ecological indicators oscillate beyond the second threshold and the structural deviation trend is within a stable range, the distribution correction mapping table in the environmental protection regulation model is activated.

[0031] Based on the distribution rules in the distribution correction mapping table, generate the distribution correction instruction for the material selection device;

[0032] If both the structural deviation trend and the ecological indicator oscillation exceed the limit, the intensity correction command generated by the energy consumption control model will be executed first, and the command of the environmental protection control model will be suspended until the energy consumption adjustment is completed.

[0033] Preferably, the step of injecting adjustment instructions into the execution process and making corrections includes:

[0034] According to the gradient value in the intensity correction instruction, adjust the real-time intensity of each area in the energy consumption device step by step;

[0035] After each adjustment, feedback data of the building structure is collected and the deviation is compared with the predicted value of the strength compensation function;

[0036] If the deviation decreases, continue the current adjustment direction until the target structure is reached;

[0037] If the deviation increases, the adjustment direction is reversed and the parameter update of the intensity compensation function is triggered again;

[0038] Based on the distribution parameters in the distribution correction instruction, the selection speed of the material selection device in different sections is dynamically adjusted.

[0039] During the distribution adjustment period, ecological indicators are monitored in real time using environmental sensors, and the coefficients in the distribution correction mapping table are updated based on the monitoring results.

[0040] Preferably, the method further includes a feedback calibration phase after the adjustment instruction is executed, the feedback calibration phase including:

[0041] Collect final structural data and ecological indicator maps after the completion of green building projects;

[0042] The final structural data is compared with the prediction range of the energy consumption control model to generate an energy consumption model error signal;

[0043] By performing overlap analysis between the ecological indicator map and the expected template of the environmental protection regulation model, an error signal for the environmental protection model is generated.

[0044] Adjust the gradient compensation rule in the energy consumption control model based on the systematic error component in the energy consumption model error signal;

[0045] Based on the random error component in the environmental protection model error signal, optimize the material distribution correction coefficient in the environmental protection control model;

[0046] The updated gradient compensation rules and material distribution correction coefficients are synchronized to the historical database of the initial parameter set.

[0047] Preferably, the process of adjusting the gradient compensation rule and the material distribution correction coefficient specifically includes:

[0048] Identify the stable deviation component in the energy consumption model error signal and calculate the compensation amount using the moving average method;

[0049] Adjust the baseline gradient in the intensity compensation function according to the compensation amount;

[0050] Identify high-frequency fluctuation components in the error signal of the environmental protection model and extract effective correction quantities through filtering techniques;

[0051] Adjust the distribution weights in the distribution correction mapping table according to the effective correction amount;

[0052] Replace the original parameters with the updated intensity compensation function and distribution correction mapping table.

[0053] Preferably, the method further includes a preprocessing phase before the project begins, wherein the preprocessing phase is performed as follows:

[0054] The climate type code of the target environment is parsed, the climate symmetry identifier and meteorological element code are extracted, and a climate feature vector is generated.

[0055] The climate feature vector is input into a pre-stored environmental characteristic database for multi-dimensional similarity search, and a candidate template set with a similarity to the current climate type higher than a threshold is selected.

[0056] For each candidate template, extract the average structural deviation and ecological indicator compliance rate from its historical records, and calculate the comprehensive effectiveness score;

[0057] The candidate templates are sorted according to the scores, and the template with the highest score is selected as the benchmark energy intensity template.

[0058] Extract the set of material distribution curves associated with the benchmark energy intensity template from the database;

[0059] For each curve, the consistency between distribution and intensity gradient is checked, and abnormal curves with abrupt changes or conflicts are removed to form an optimized set of material distribution curves.

[0060] Based on the historical stability index of the optimized curve set, the curve with the smallest fluctuation is selected as the benchmark material distribution curve.

[0061] The baseline energy intensity template and the baseline material distribution curve are time-aligned to generate a baseline configuration for the initial parameter set.

[0062] Preferably, the steps for verifying the compatibility of the material distribution curve include:

[0063] Extract individual curve data from the curve set and obtain the intensity gradient sequence corresponding to time in the benchmark energy intensity template;

[0064] Based on the time points of the intensity gradient sequence, collaborative timestamps are marked on the curve to generate a marked distribution curve;

[0065] Traverse the marked curves and detect whether the slope of the distribution change within adjacent time intervals exceeds the limit to identify abnormal intervals;

[0066] When an abnormal interval is detected, the gradient value at the corresponding time point in the intensity gradient sequence is traced back to determine whether the gradient fluctuation direction conflicts with the direction of the sudden change in distribution.

[0067] If the conflict intensity exceeds the limit, it is marked as a conflict area, and its starting position and duration are recorded.

[0068] Based on the location and duration of the conflict area, a correction interval is defined on the curve, and a smooth replacement segment is generated based on historical data.

[0069] Insert the replacement segment into the correction interval to generate the optimized curve, and delete the original abnormal data points;

[0070] After completing the replacement operation on all curves, perform an integrity check and remove any uncorrected residual curves.

[0071] The obtained curves are merged into an optimized material distribution curve set.

[0072] Preferably, the method further includes a real-time optimization iteration phase after generating the adjustment instructions, the real-time optimization iteration phase including executing:

[0073] Continuously collect new environmental monitoring data during project implementation and update the initial parameter set;

[0074] Based on the updated parameter set, the applicability of the energy consumption control model and the environmental protection control model are re-verified.

[0075] When the verification fails, the model retraining process is initiated to regenerate the intensity compensation function and distribution correction mapping table using the latest data.

[0076] The retrained model parameters are integrated into the tuning instructions to form a closed-loop optimization mechanism.

[0077] Compared with the prior art, the beneficial effects of the present invention are:

[0078] From the perspective of parameter acquisition and integration, this method acquires real-time monitoring data of the target building environment to construct an initial parameter set encompassing energy intensity gradient sequences and material distribution curves. This breaks through the limitations of traditional design, which relies solely on single energy consumption data or static material indicators. The energy intensity gradient sequence can comprehensively present the energy consumption variation patterns of different building structural areas and time periods, while the material distribution curves can clearly reflect the layout of materials within the overall building. The combination of these two elements allows the initial parameter set to more comprehensively depict the actual situation of the target building environment, providing richer and more accurate basic data for the development of subsequent control models, and making the model construction more closely aligned with the actual operational needs of the building.

[0079] At the level of regulatory model development, this method constructs dedicated regulatory models for the two core dimensions of energy consumption and environmental protection, with model design closely aligned with the characteristics of the target environment. The energy consumption regulatory model focuses on the dynamic relationship between energy intensity and building structure, and can provide real-time feedback on the fluctuation trend of energy intensity based on changes in building structure (such as adjustments to the thermal insulation performance of doors and windows, and optimization of ventilation systems), avoiding the problem of regulatory failure in traditional static models when the environment changes. The environmental protection regulatory model specifically expresses the nonlinear relationship between material distribution and ecological indicators, and can accurately capture the impact of material layout adjustments (such as the increase of high-environmental-friendly materials in key areas and the replacement of high-energy-consuming materials) on ecological indicators. This allows material selection and layout to no longer be limited to a single environmental protection indicator, but to truly adapt to the ecological needs of the target environment. For example, in ecologically sensitive areas, the model can optimize material distribution to reduce the interference of building material use on the surrounding environment.

[0080] In the model application and execution adjustment phase, this method introduces a dynamic model selection mechanism, which can flexibly switch between energy consumption control models and environmental protection control models based on real-time monitoring of structural deviation data and ecological indicator fluctuation data during the design phase. This dynamic selection method avoids the lag of traditional manual judgment. When structural deviations occur in the building (such as insufficient insulation layer thickness due to construction), the energy consumption control model can be quickly invoked to output strength gradient adjustment commands to promptly compensate for energy losses. When ecological indicators fluctuate (such as a temporary increase in regional carbon emissions), the environmental protection control model can be quickly activated to optimize material distribution curves, reducing the building's additional burden on the environment. At the same time, adjustment commands are directly injected into the design execution process, enabling real-time correction of energy consumption devices or material selection devices, ensuring that control measures can be implemented quickly, avoiding a disconnect between the design scheme and actual implementation, and reducing the loss of energy-saving and environmental protection effects due to adjustment lags. Attached Figure Description

[0081] Figure 1 This is a schematic diagram illustrating the working principle of the energy-saving and environmentally friendly green building design method described in this invention.

[0082] Figure 2 A flowchart for constructing a refined initial parameter set;

[0083] Figure 3 This is a flowchart for the feedback calibration phase. Detailed Implementation

[0084] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0085] Please see Figure 1 This invention provides an energy-saving and environmentally friendly green building design method. This method achieves synergistic optimization of building energy consumption and environmental performance by integrating real-time monitoring, dynamic modeling, and adaptive adjustment mechanisms. The overall process of the green building design method is based on continuous monitoring of the target building environment. The construction of the initial parameter set includes an energy intensity gradient sequence and material distribution curves. The energy intensity gradient sequence reflects the energy consumption variation trend in different areas of the building, and the material distribution curves describe the configuration pattern of environmentally friendly materials in the building structure. The development of the energy consumption control model relies on the energy intensity gradient sequence, which establishes the mathematical relationship between energy consumption intensity and building structural parameters. The formation of the environmental protection control model is based on the material distribution curves, which captures the complex interaction between material selection and ecological indicators.

[0086] In the actual operation of green building design, real-time collected structural deviation data and ecological index fluctuation data are used to trigger model selection logic. The energy consumption control model or environmental protection control model is dynamically activated to generate adjustment instructions. The adjustment instructions directly affect the control system of the energy consumption device or the configuration process of the material selection device, realizing real-time correction of the intensity gradient or distribution curve, thereby improving the overall energy efficiency and environmental performance of green buildings.

[0087] Example 1:

[0088] See appendix Figure 2 An environmental sensor network is deployed in the boundary areas and interior spaces of the target building environment. The sensors include temperature, humidity, and light intensity sensors, which continuously record raw meteorological parameter data during the planning period in a synchronous acquisition mode. Temperature change curves are analyzed using time-series data to extract feature points, and humidity distribution data is used to generate a two-dimensional distribution map using spatial interpolation algorithms. The raw data is transmitted to the central processing unit via a communication module for preprocessing. The preprocessing stage eliminates sensor noise and compensates for transmission delays, resulting in standardized temperature change curves and humidity distribution datasets. An extreme value region identification algorithm based on the temperature change curves is launched. Extreme value regions are defined as continuous spatiotemporal units where the temperature gradient exceeds a set threshold. The algorithm marks the geometric center coordinates of high-temperature and low-temperature extreme value regions. Based on the spatial distribution of extreme value regions, the target environment is divided into multiple sub-regions. Sub-region boundaries are automatically generated using the Voronoi diagram algorithm, and each sub-region is assigned an independent identifier code. The initial energy consumption gradient range is set based on the historical energy consumption baseline of the sub-regions. Material distribution reference values ​​are derived from the default configuration of the building materials database. The sub-region management module establishes a spatial index to support rapid retrieval.

[0089] The pre-configured database is connected to a distributed storage system. Database queries retrieve corresponding historical energy intensity compensation parameters based on sub-region identifiers. These historical energy intensity compensation parameters contain a time-stamped sequence of energy consumption corrections, with data extraction covering records from the three most recent full years. Time-domain filtering employs a Butterworth low-pass filter, with the filter cutoff frequency dynamically adjusted according to seasonal cycle characteristics. The filtered output eliminates short-term fluctuations while preserving long-term trends. The optimized energy intensity gradient sequence is normalized to a standard format, and the sequence data is stored in the parameter set cache according to sub-region numbers. Material distribution adjustment coefficients are obtained in batches from the database association table. The coefficient matrix contains spatial weight values ​​for different material types, and frequency domain smoothing uses a combination of Fourier transform and inverse transform algorithms. Frequency domain analysis separates the fundamental and harmonic components. The harmonic suppression threshold is set based on the material characteristic curves, and the smoothed coefficient matrix is ​​remapped to a spatial coordinate system. The optimized material distribution curve generation process includes curve fitting and outlier removal. The fitting algorithm uses the least squares method to ensure curve continuity, and the optimized material distribution curve for each sub-region is stored independently with a version marker. The initial parameter set construction module integrates data from all sub-regions. The integration process includes data alignment and format verification. The energy intensity gradient sequence and material distribution curve are precisely synchronized using timestamps. Parameter set integrity verification is performed using hash value validation. The final generated initial parameter set is written to the project database and a configuration log is generated.

[0090] During the development phase of the energy consumption regulation model, optimized energy intensity gradient sequences are loaded into the model training platform. A multilayer perceptron neural network structure is chosen for the regression model, with the number of input layer nodes corresponding to the time dimension of the gradient sequence. Hidden layer weights are initialized using the Xavier method, and the training iterations are set to 1000, with each iteration including both forward and backpropagation phases. The ReLU function is used to calculate hidden layer activations during forward propagation, and an adaptive moment estimation algorithm is used for backpropagation error correction. The weight update step size is dynamically adjusted based on the gradient magnitude. The intensity compensation function is extracted from the output layer of the trained regression model; its expression includes a piecewise linear approximation formula, and the compensation function parameters are encapsulated as a callable software module.

[0091] The environmental regulation model training process loads optimized material distribution curves into the clustering analysis engine. The clustering model uses the k-means algorithm, and the number of cluster centers is automatically determined based on the number of material categories. The backpropagation algorithm optimizes the movement trajectory of cluster centers, and the sample distance calculation uses the Euclidean metric formula. The iteration termination condition is set to the center point movement distance being less than a threshold. The distribution correction mapping table generation process includes cluster labeling and mapping rule definition. Each cluster generates a corresponding material distribution correction parameter table, and the mapping table data structure is stored in key-value pair format. The model convergence detection module monitors the loss function curves of the regression model and the clustering model. The convergence criterion is that the rate of change of the loss value is lower than a set value for 100 consecutive iterations. The feature vector extraction operation extracts slope and intercept parameters from the intensity compensation function and density distribution feature values ​​from the distribution correction mapping table. After feature vector normalization, similarity calculation is performed with real-time meteorological parameters using the cosine similarity algorithm. The matching results are compared with a preset threshold to generate a verification report. When the matching degree falls below the set standard, the training iteration adjustment module automatically increases the iteration count of the regression and clustering models, with an increment step size of 50% of the original number. The retraining process employs an incremental learning strategy to retain the original weights. Once the feature vectors meet the applicable conditions, the model parameters are permanently stored in the model library and the version number is updated.

[0092] The data acquisition frequency of the environmental sensor network is dynamically adjusted according to the rate of meteorological change, with the acquisition interval shortened to 5 minutes during high-temperature periods and extended to 15 minutes during low-temperature periods. The data transmission protocol uses MQTT to ensure real-time performance, and the data buffer of the central processing unit implements a circular buffer management strategy to prevent data overflow. The extreme value detection algorithm for temperature change curves introduces a sliding window mechanism, with the window size adaptively adjusted according to the daily average temperature change amplitude to avoid missing transient extreme points. A manual correction interface is introduced during sub-region segmentation, allowing designers to fine-tune the automatically generated boundaries; fine-tuning data is recorded in the operation log for audit traceability. The retrieval of historical energy intensity compensation parameters supports multi-condition combined queries, with query conditions including auxiliary parameters such as season type, building occupancy rate, and equipment operating status. The cutoff frequency optimization of time-domain filtering is based on spectrum analysis results; a lower cutoff frequency is used for summer data to eliminate air conditioning load fluctuations, while a higher cutoff frequency is used for winter data to preserve heating characteristics.

[0093] The frequency domain smoothing of the material distribution adjustment coefficients incorporates wavelet transform as an alternative. The algorithm automatically switches when the Fourier transform exhibits Gibbs free energy. Smoothing quality is assessed through signal-to-noise ratio calculation. The goodness-of-fit test of the optimized material distribution curve uses the R-squared index. Unacceptable curves trigger a refit process, with fitting parameters iteratively optimized until accuracy requirements are met. An early stopping mechanism is introduced during regression model training to prevent overfitting. Training automatically terminates when the validation set error continuously increases. The model structure supports dynamically expanding the number of hidden layer nodes. The initial centroid selection for the clustering model is optimized using the k-means++ algorithm to avoid getting trapped in local optima. The number of clusters is verified using silhouette coefficients. A weighted strategy is introduced for eigenvector matching degree calculation, assigning differentiated weights to different feature parameters. The weight values ​​are set based on feature importance analysis results. The model retraining process employs a rolling time window strategy, using only the most recent data to ensure model timeliness. Historical data is archived to a knowledge base for trend analysis.

[0094] Example 2:

[0095] Monitoring of structural deviation trends is achieved through a network of strain sensors deployed at the building's load-bearing nodes. These sensors continuously collect structural deformation data at a 100Hz sampling frequency, and the data stream is digitally filtered to extract the trend component. Monitoring of the oscillation range of ecological indicators relies on a multispectral environmental sensor array. This array measures the concentration of particulate matter, volatile organic compounds, and noise levels in the air, and the measured values ​​are statistically analyzed for oscillation amplitude within a time window. Monitoring data is transmitted to the decision-making center via industrial Ethernet. The data preprocessing module calculates the first derivative of the structural deviation trend and the moving variance of the ecological indicator oscillation range. A threshold comparator array is installed within the decision-making center. The first threshold is preset to 80% of the allowable structural deviation value, and the second threshold is set to 75% of the safe range for ecological indicators. The threshold comparators receive the structural deviation trend signal and the ecological indicator oscillation range signal in real time. After analog-to-digital conversion, the signals are compared with the threshold values. When the structural deviation trend value exceeds the first threshold and the ecological indicator oscillation range value is within a stable range, the comparator outputs a high-level signal to trigger the energy consumption control model. In the energy consumption control model, the intensity compensation function activates the function calculation unit after receiving the start signal. The function calculation unit queries the gradient adjustment rule base based on the real-time deviation value. The gradient adjustment rule base stores the correction coefficients corresponding to different deviation levels. The intensity correction instruction generation module encodes the correction coefficients into control instructions that the equipment can recognize. The control instructions include the target intensity value, adjustment rate, and effective timestamp.

[0096] When the oscillation range of ecological indicators exceeds the second threshold and the structural deviation trend is within a stable range, the threshold comparator outputs another high-level signal to trigger the environmental control model. The distribution correction mapping table in the environmental control model initiates a pattern matching algorithm, which calculates the similarity between the current ecological indicator oscillation pattern and historical patterns in the mapping table. The distribution correction instruction generation module extracts distribution rule parameters based on the matching results. These parameters include material replacement priority and distribution density adjustment. The instruction encoding uses JSON format to encapsulate the device operation sequence. In complex situations where both the structural deviation trend and the ecological indicator oscillation range exceed the threshold simultaneously, a priority arbitrator intervenes in the instruction scheduling process. The priority arbitrator uses a fixed priority strategy, with the intensity correction instruction generated by the energy consumption control model receiving the highest priority, and the distribution correction instruction generated by the environmental control model entering the waiting queue. A circular buffer is created in the instruction temporary storage area to store instructions to be executed. The buffer management strategy uses the least recently used algorithm, and the energy consumption adjustment completion signal triggers the environmental instruction dequeue operation.

[0097] The execution of strength correction commands is achieved through the building equipment management system. The execution engine of the building equipment management system parses the gradient values ​​in the commands, decomposing them into multiple adjustment steps. The control interface of the energy consumption device receives the adjustment step sequence, with the step interval dynamically set according to the equipment response characteristics. After each step is executed, a feedback data acquisition loop is triggered. Feedback data from the building structure is acquired through a fiber optic sensor network. The data includes stress distribution maps and vibration spectrum characteristics. The feedback data and the predicted value of the strength compensation function are differentially calculated in a data comparator. A difference value less than zero indicates a reduction in deviation, the control logic maintains the current adjustment direction, and the actuator continues to operate at the original step size. A difference value greater than zero triggers an adjustment direction reversal mechanism. The direction reversal signal is simultaneously sent to the strength compensation function parameter update module, which recalibrates the function coefficients. The material selection device's distribution correction command execution process employs an adaptive control strategy. The distribution parameter conversion device maps the distribution parameters in the commands into equipment control signals, which are transmitted to the material conveying robotic arm via a fieldbus. The robotic arm controller adjusts the grasping frequency and movement trajectory according to the distribution parameters, and the selection speed of different sections is infinitely adjustable through a variable frequency speed control mechanism.

[0098] During distribution adjustment, environmental monitoring incorporates a closed-loop feedback mechanism. Real-time environmental sensor data is input into the distribution correction mapping table update algorithm, which calculates the weighted average of old and new parameters. The mapping table version management module records the history of each parameter change, and the parameter rollback function automatically restores the previous stable version if the adjustment effect is unsatisfactory. A complete audit trail is established throughout the entire instruction execution process, with operation logs generated for each adjustment step. Log data is used for subsequent performance analysis and optimization. The algorithm for monitoring structural deviation trends incorporates multi-scale analysis technology. Short-term trends use a sliding window to calculate instantaneous derivatives, while long-term trends use exponential smoothing to extract baseline components. The statistical analysis of the oscillation range of ecological indicators incorporates box plot analysis to automatically identify abnormal fluctuations and eliminate interfering data points. The threshold comparator is designed with a hysteresis comparison circuit to prevent critical state jitter. Input validity verification is added to the function calculation process after the energy consumption control model starts. Invalid inputs trigger a backup rule base switching mechanism, and the gradient adjustment rule base supports online hot updates without downtime maintenance. The pattern matching algorithm for the environmental control model uses an improved dynamic time warping algorithm, supporting similarity calculations for time series of different lengths. Fuzzy logic reasoning is incorporated into the generation of distribution rule parameters. The priority arbiter supports manual intervention under extreme conditions, allowing operators to adjust the command execution order via a human-machine interface. A command buffer provides persistent data storage to prevent data loss during power outages. The execution of intensity correction commands employs a progressive safety strategy, with the maximum single-step adjustment constrained by a safety limiter. Equipment response delay is incorporated into the step interval calculation model.

[0099] The control system of the material selection device incorporates a collision avoidance detection algorithm, performing real-time collision detection on the robotic arm's motion trajectory and adding acceleration smoothing processing during the speed adjustment process. Closed-loop feedback of environmental sensor data employs predictive correction technology, fusing current monitoring data with predicted values, and dynamically adjusting the mapping table update frequency based on system load. The entire implementation system is designed with a redundant backup mechanism, with key modules using a dual-machine hot standby architecture, and the fault detection module automatically triggering a switchover to the backup system.

[0100] Example 3:

[0101] See appendix Figure 3The feedback calibration phase is initiated within a specific time window after the green building project is completed. The data acquisition system collects the final structural dataset and ecological indicator map set throughout the project's entire lifecycle. The final structural dataset includes the stress distribution matrix, vibration frequency spectrum, and thermal performance parameter table of the building's load-bearing structure. The ecological indicator map set covers air quality index cloud maps, acoustic environment contour maps, and light pollution distribution matrices. The data comparison module performs a multi-dimensional comparison between the final structural data and the prediction range of the energy consumption control model. The comparison algorithm uses spatial registration technology to align actual measurement points with predicted grid points, generating an energy consumption model error signal with spatiotemporal labels. The energy consumption model error signal contains systematic error components and random error components. The systematic error component manifests as a persistent deviation vector, while the random error component exhibits high-frequency fluctuation characteristics.

[0102] The ecological indicator map and the expected template of the environmental protection regulation model are subjected to overlap analysis. The overlap analysis adopts a convolutional neural network algorithm from the field of image processing, in which the convolutional kernel slides in the map space to calculate local similarity. The overlap analysis outputs the environmental protection model error signal. The random error component in the environmental protection model error signal is separated into high-frequency components by wavelet transform, and the systematic error component is identified by the baseline offset through a trend extraction algorithm. The error signal decomposition module inputs the stable deviation component in the energy consumption model error signal into a moving average filter. The window size of the moving average filter is automatically adjusted according to the periodicity of the error signal, and the filtered output is the smoothed compensation amount.

[0103] The gradient compensation rule in the energy consumption control model is adjusted by adjusting the compensation amount. The modification of the gradient compensation rule is achieved through a parameter optimization algorithm. The formula for calculating the adjustment amount ΔB of the baseline gradient is as follows:

[0104]

[0105] Where α represents the learning rate coefficient, n represents the moving average window size, and ε sys (t) represents the system error value at time t, and ∑ represents the summation operation. The baseline gradient update module adds the adjustment to the original baseline value to generate a new version of the gradient compensation rule.

[0106] The high-frequency fluctuation components in the environmental protection model error signal are processed by digital filtering technology. A Butterworth band-stop filter removes noise interference in specific frequency bands, and the effective correction amount is extracted using an envelope detection algorithm to extract amplitude features. The distribution weights in the distribution correction mapping table are iteratively updated based on the effective correction amount. The weight adjustment uses the gradient descent method to minimize the error function, and the updated distribution weights reconfigure the query parameters of the mapping table. The parameter replacement module replaces the original parameters with the updated intensity compensation function and the distribution correction mapping table. The replacement process uses a transaction mechanism to ensure data consistency, and a version control system records the parameter change history. The final structure dataset is acquired through a distributed sensor network, which includes fiber optic strain sensors, microelectromechanical accelerometers, and infrared thermal imager arrays. A data fusion algorithm unifies the heterogeneous sensor data into a standard coordinate system. The prediction range of the energy consumption control model is generated by the model inference engine. The inference engine loads a trained neural network model for forward computation, and the prediction results are stored in a three-dimensional grid. Uncertainty quantification technology is introduced in the generation of the error signal. Each error value is marked with a confidence interval, and principal component analysis is used to separate correlation features for the identification of systematic error components. Ecological indicator maps were collected using a combined system of multispectral remote sensing equipment and ground monitoring stations. The remote sensing equipment provided continuous spatial monitoring data, while the ground monitoring stations provided precise, fixed-point measurements. The expected template for the environmental regulation model was generated through numerical simulation, which considered seasonal variations and the impact of human activities. The template data employed a hierarchical storage structure to support rapid retrieval. The convolutional neural network used in the overlap analysis process employed a pre-trained semantic segmentation model, with the feature extraction layer using a ResNet architecture, and similarity calculation using a cosine distance metric.

[0107] The window size optimization of the moving average filter is based on spectral analysis results. The window length is matched with the main periodic components of the error signal, and the filtering effect is evaluated by the signal-to-noise ratio improvement metric. The learning rate coefficient in the baseline gradient adjustment formula adopts an adaptive adjustment strategy, with an initial value set at 0.01, dynamically adjusted based on historical adjustment results. A momentum term is introduced during the distributed weight update process to prevent local optima; the momentum coefficient is set to 0.9 to accelerate convergence, and weight pruning techniques prevent gradient explosion. A version control system records metadata for each parameter change, including modification time, operator, and reason for change. The parameter rollback function supports quick restoration to any historical version. A quality assurance system is established throughout the feedback calibration phase, with each processing step outputting a quality report, and abnormal data triggering a reprocessing flow. Long-term tracking analysis of the error signal reveals systematic deviation patterns, which are categorized using clustering algorithms. The categorization results are used to guide model structure optimization. The overlap analysis of the ecological indicator map introduces multi-scale feature matching technology. The matching algorithm calculates similarity on a spatial pyramid structure, improving matching accuracy. Regularization constraints are introduced during parameter updates to prevent overfitting. The regularization coefficients are automatically adjusted based on the dataset size, and cross-validation is used to evaluate the generalization ability of the parameters.

[0108] The feedback calibration phase works in tandem with the project acceptance process, and calibration results are a crucial component in generating the acceptance report. Report data is stored using blockchain technology to ensure its immutability. Intermediate data generated during calibration is archived in a knowledge base, which is indexed to support rapid retrieval of similar projects. Data mining techniques uncover optimization patterns from historical calibration records. The system implements an automated calibration pipeline, triggering calibration tasks periodically, and automatically pushing calibration results to relevant system modules. A visual configuration interface is provided for manual intervention. The effectiveness of the feedback calibration phase is continuously tracked through an indicator monitoring system, including model accuracy improvement rate, parameter convergence speed, and system stability coefficient. Monitoring data is displayed in real-time in the management dashboard. An anomaly detection module monitors sudden deviations during calibration, and anomaly point analysis helps identify potential system problems, with early warning information being pushed to maintenance personnel at different levels. Calibration logs record the complete operation trajectory, and log analysis tools support performance bottleneck location and continuous optimization of calibration algorithm efficiency. The technical implementation of the feedback calibration phase adopts a microservice architecture, with each processing module deployed and running independently. Services communicate asynchronously via message queues, and containerization technology ensures environmental consistency. The data persistence layer uses a time-series database to store monitoring data and a relational database to store metadata, with a caching mechanism improving data access speed. The system's security design includes multiple layers of protection such as identity authentication, access control, and data encryption. The audit function records all data access operations, meeting the requirements of information security standards.

[0109] Example 4:

[0110] The preprocessing phase begins immediately after the green building project is approved. The climate data analysis module downloads historical climate data for the target area from the national meteorological database. This historical climate data includes raw observation records of temperature, humidity, wind speed, and sunshine duration over the past thirty years. A climate type code generation algorithm performs cluster analysis on the raw data. Cluster features include annual temperature range, seasonal precipitation distribution type, and prevailing wind frequency. Each cluster center is assigned a unique climate type code. Climate symmetry identifiers are derived by calculating the coefficient of variation of meteorological elements between seasons. Meteorological element coding adopts the international standard WMO coding rules, and the climate feature vector consists of a 128-dimensional floating-point vector space. The environmental characteristic database is deployed in a distributed storage system. The database index structure uses a B+ tree to optimize query speed, and the multi-dimensional similarity search algorithm uses locality-sensitive hashing (LSH) technology to accelerate computation. The similarity threshold is set to 0.85. The candidate template set generation process includes deduplication and validity verification. For each candidate template, the mean structural deviation and ecological indicator compliance rate are extracted from its project file. The comprehensive performance score calculation formula includes a weighting mechanism, with structural deviation weighted at 0.6 and ecological indicators weighted at 0.4. The score results are normalized. The template sorting module arranges candidate templates in descending order of score, and the highest-scoring template activates the benchmark energy consumption intensity template loading process. A correlation query retrieves the corresponding material distribution curve set from the material library. The coordination test initiates a data preprocessing procedure, standardizing curve data to eliminate the influence of dimensions. A time alignment algorithm synchronizes the time-series data of the material distribution curves and energy consumption intensity templates to a unified time axis. Coordination timestamps are placed at curve inflection points, and time intervals are divided using equally spaced sliding windows. The distribution change slope is calculated using the central difference method. Anomaly interval detection sets a slope threshold of ±15% / hour. The conflict judgment module compares the angle between the gradient fluctuation direction and the distribution change direction; an angle exceeding 90 degrees is considered a directional conflict. Conflict areas are marked with a red warning indicator. The smooth replacement segment generation uses a cubic spline interpolation algorithm, with interpolation nodes taken from historical data two hours before and after the conflict area. The integrity check verifies the continuity and boundary conditions of the curves. Residual curve identification rules are based on data integrity indicators, and data consistency checks are performed before outputting the optimized material distribution curve set. The baseline configuration generation module performs time alignment operations with millisecond-level accuracy, and the baseline configuration of the initial parameter set is written to the project configuration file. Outlier handling during climate data parsing employs box plots to remove abnormal meteorological records exceeding 1.5 times the interquartile range, and missing data is filled using spatiotemporal kriging interpolation. Climate type code generation incorporates a machine learning classification model, trained using labeled data from national meteorological stations. Classification features include 12 key indicators such as annual precipitation, accumulated temperature, and frost days. Climate symmetry identification calculation introduces the concept of information entropy; the entropy value reflects the uniformity of climate element distribution, and meteorological element coding follows the WMO code table for digital conversion.The similarity search of the environmental characteristic database implements a parallel computing architecture. The query task is decomposed into multiple subtasks and distributed to computing nodes. The search results aggregation adopts a max-heap data structure. Deduplication of the candidate template set is based on template fingerprint information. Fingerprint generation uses the MD5 hash algorithm, and validity verification checks the integrity and logical consistency of the template data. The comprehensive performance score calculation incorporates a time decay factor, assigning higher weight to recent project data, and the score results are standardized using Z-score. Curve preprocessing for the consistency test includes outlier removal and smoothing. The smoothing algorithm uses a Savitzky-Golay filter to preserve curve features. A time alignment algorithm addresses the synchronization problem of data with different sampling frequencies, and linear interpolation is used to ensure data accuracy. The identification of collaborative timestamps combines curvature change features, with high curvature points used as key marker positions. The calculation of the distribution change slope incorporates robust estimation techniques to avoid interference from local jitter in judging the overall trend.

[0111] Conflict zone determination employs a multi-indicator comprehensive decision-making process, considering three dimensions: mutation magnitude, duration, and impact range. The generation of smooth replacement segments utilizes constraint optimization methods to ensure curve smoothness. Integrity checks are implemented through an automated testing process, with test cases covering various boundary conditions. The optimized curve set is then visually verified. After the baseline configuration is generated, stress testing is performed to simulate parameter stability under extreme climatic conditions, and a backup mechanism is configured to ensure data security.

[0112] Principal component analysis is used for dimensionality reduction of climate feature vectors, retaining principal components with a contribution rate exceeding 95%. An improved cosine similarity algorithm is used for vector similarity calculation. The environmental characteristic database index update mechanism implements incremental updates, automatically triggering index rebuilding for new project data. Database partitioning based on climate type improves query efficiency. The candidate template set selection criteria support multiple combinations, including auxiliary parameters such as building type, building area, and structural form. Template metadata records basic project information and implementation time. The algorithm optimization for consistency verification incorporates dynamic programming to improve the efficiency of long-sequence data processing, and the sensitivity parameters for conflict detection support tiered settings. The generation of smooth substitution segments considers material characteristic constraints, with differentiated smoothing parameters for different material types. Substitution operations are logged for audit traceability. Integrity checks employ a multi-round iterative strategy: the first round checks basic data attributes, the second round checks logical relationships, and the third round checks business rules. The generation of the baseline configuration includes version control information; each modification generates a configuration snapshot, and the rollback function supports rapid restoration of historical versions. A quality assurance system is established for the preprocessing stage, with quality checkpoints set at each processing stage. The abnormal data handling process includes a manual review step. The system performance monitoring tracks processing progress in real time, the resource scheduling algorithm optimizes the allocation of computing resources, and task priorities are dynamically adjusted according to the urgency of the project. Preprocessing results generate detailed technical reports, including data processing methods, parameter selection criteria, and quality assessment results, with the report format conforming to patent examination standards.

[0113] The preprocessing stage employs a modular architecture, with each processing module independently encapsulated as a microservice, and services communicating via RESTful APIs. The data persistence layer utilizes a hybrid storage solution, storing hot data in an in-memory database and archiving cold data to an object storage system. System security features include encrypted data transmission, access control, and operation log auditing. System reliability is ensured through load balancing and failover mechanisms. The preprocessing system establishes standard data interfaces with subsequent design stages, enabling seamless data flow throughout the entire process.

[0114] Example 5:

[0115] Taking a commercial complex project in Pudong New Area, Shanghai as an example, the environmental monitoring sensor network deployed on-site collects temperature, humidity, and light intensity data once per minute. The data is transmitted to the central processing system via a 5G network. New environmental monitoring data includes the vertical temperature gradient distribution from the third basement level to the twentieth floor above ground. This temperature gradient data is collected through temperature measurement chains installed in elevator shafts, with each measurement node spaced five meters apart. The initial parameter set update module refreshes the data every hour, retaining a sliding window of data from the most recent 72 hours. The energy consumption intensity gradient sequence is recalculated using a weighted average algorithm, and spatiotemporal interpolation technology is introduced to update the material distribution curve. The model applicability verification program is automatically triggered after each parameter set update. The verification process analyzes the deviation between real-time monitoring data and model predictions. Taking the second quarter of 2023 data from the Shanghai project as an example, the verification program found that the energy consumption control model's prediction deviation on high-temperature days in summer exceeded the allowable range, with the deviation mainly concentrated in the period from 2 PM to 4 PM. The model retraining process is immediately initiated. The training dataset contains complete monitoring records from the most recent three months, and data standardization eliminates the influence of the diurnal cycle. The intensity compensation function is regenerated using an improved random forest algorithm. The algorithm's input features include twelve dimensions such as temperature, humidity, and population density. Ten-fold cross-validation is performed during the function training process. The distribution correction mapping table is reconstructed using a fuzzy clustering method. The number of clusters is automatically determined based on the silhouette coefficient, and the mapping table output undergoes a consistency check.

[0116] The closed-loop optimization mechanism integrates the retrained model parameters into the adjustment command generation pipeline, converting the command format into a Modbus protocol recognizable by the building equipment management system. In the Shanghai project example, adjustment commands are sent to the air conditioning unit controller via an industrial IoT gateway. The controller adjusts the chiller unit's outlet water temperature according to the new strength compensation function. The material selection system receives the updated distribution correction mapping table, which guides the automated warehousing system to adjust the order of insulation material releases. The release priority is dynamically adjusted based on real-time outdoor temperature. The acquisition of newly added environmental monitoring data achieves a fully automated process. The sensor network adopts a self-organizing mesh topology, automatically detecting node failures and activating backup transmission paths. The data preprocessing stage includes outlier filtering and data imputation. Outlier identification uses the isolated forest algorithm, and data imputation uses a multiple interpolation method. The initial parameter set update strategy supports manual adjustment of the update frequency, and can trigger immediate updates in emergencies. The update log records detailed parameters for each change.

[0117] The model applicability verification design incorporates a multi-level alarm mechanism, with deviation levels categorized into three levels: Attention, Warning, and Severe, each triggering a differentiated processing procedure. During the verification process in the Shanghai project, it was discovered that the energy consumption prediction for the southeast-facing area consistently deviated, ultimately leading to an error in the input of the solar radiation coefficient for the glass curtain wall. The model retraining process introduced incremental learning technology, which retains existing knowledge structures while incorporating new data features, improving training efficiency by 40%. After comparing multiple options, the random forest algorithm was ultimately selected as the optimal choice for handling nonlinear relationships in the intensity compensation function algorithm selection. The fuzzy clustering process of the distribution correction mapping table uses a dynamic threshold, which automatically adjusts based on data distribution density. The clustering results are visualized for manual review. A closed-loop optimization mechanism establishes a two-way communication channel, with real-time feedback from equipment execution transmitted back to the optimization system. This feedback data is used to evaluate the adjustment effect. In the Shanghai project, after the air conditioning system adjustment command was executed, the system monitored the power change curve in real time, comparing it with the expected energy-saving effect. A complete performance monitoring system was constructed during the real-time optimization iteration phase, with monitoring indicators including model accuracy, command response latency, and system stability coefficient. Monitoring data is displayed in real time through dashboards, supporting multi-dimensional drill-down analysis. Historical data is stored for the entire project lifecycle. The anomaly handling mechanism includes an automatic rollback function; when optimization results do not meet expectations, the system automatically reverts to the previous stable version, and detailed logs are recorded for the rollback operation.

[0118] Taking a project implementation as an example, a sudden temperature rise occurred on a day during the spring transition season. The real-time optimization and iteration system completed the entire process of data collection, model retraining, and command issuance within two hours. The adjustment command raised the air conditioning temperature setpoint in public areas by two degrees Celsius and automatically adjusted the lighting brightness in corridor areas to energy-saving mode, with the system adaptively adjusting to avoid energy waste. The material selection system corrected the mapping table based on the new distribution, increasing the thickness of the insulation material planned for the exterior walls by five millimeters to compensate for heat loss caused by temperature fluctuations. The technical architecture of the real-time optimization and iteration phase adopts a cloud-edge-device collaborative computing model. The cloud is responsible for complex model training, edge nodes process real-time control commands, and terminal devices execute specific operations. The data security mechanism includes triple protection: transmission encryption, storage encryption, and access control. Audit tracking records all data access and modification operations. System reliability is ensured through multiple backup mechanisms, with the primary and backup systems using a hot backup mode and a switchover time of less than three seconds.

[0119] The real-time optimization iteration phase is deeply integrated with the Building Information Modeling (BIM) system, automatically synchronizing optimization results to the BIM and updating relevant parameter settings. Maintenance personnel can view the optimization status in real-time via mobile devices and remotely adjust optimization strategy parameters; the system supports multi-user collaborative operation. Version management records the complete history of each optimization iteration, supporting status regression at any point in time and comparison of optimization effects across different versions. The implementation effectiveness of the real-time optimization iteration phase is evaluated using key performance indicators (KPIs), including energy intensity, material utilization efficiency, and system response time. Evaluation reports are automatically generated periodically, and the report data supports decision-making optimization and continuous improvement of the real-time optimization iteration algorithm. The knowledge base system accumulates optimization cases and experience data, allowing new projects to directly refer to optimization solutions for similar scenarios, improving project implementation efficiency. A data interface with the weather forecasting system is established during the real-time optimization iteration phase, with weather forecast data serving as a pre-input for model prediction, improving the foresight of adjustment instructions.

[0120] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An energy-saving and environmentally friendly green building design method, characterized in that, The method implements the following steps: Acquire real-time monitoring data of the target building environment, and construct an initial parameter set based on the real-time monitoring data. The initial parameter set includes an energy intensity gradient sequence and a material distribution curve. Based on the energy intensity gradient sequence in the initial parameter set, an energy consumption control model that is coordinated with the target environment is developed. The energy consumption control model is used to express the dynamic relationship between energy intensity and building structure. Based on the material distribution curves in the initial parameter set, an environmental protection regulation model that fits the target environment is formed. The environmental protection regulation model is used to express the nonlinear relationship between material distribution and ecological indicators. During the green building design phase, based on real-time monitored structural deviation data and ecological indicator fluctuation data, an energy consumption control model or an environmental protection control model is dynamically selected, and adjustment instructions are output through the selected model. Adjustment instructions are injected into the execution process of green building design to enable real-time correction of the intensity gradient of energy consumption devices or the distribution curve of material selection devices.

2. The energy-saving and environmentally friendly green building design method as described in claim 1, characterized in that, The steps for acquiring real-time monitoring data and constructing an initial parameter set are further refined as follows: Multiple environmental sensors are used to simultaneously collect meteorological parameters of the target environment during the planning period, including temperature change curves and humidity distribution data; Based on the extreme regions in the temperature change curve, the target environment is divided into multiple sub-regions, and an initial energy consumption gradient range and material distribution reference value are set for each sub-region. Retrieve historical energy consumption intensity compensation parameters and material distribution adjustment coefficients corresponding to each sub-region from the preset database; By performing time-domain filtering on the historical energy consumption intensity compensation parameters, the optimized energy consumption intensity gradient sequence for each sub-region is obtained. The material distribution adjustment coefficient is smoothed in the frequency domain to generate an optimized material distribution curve for each sub-region; The optimized energy intensity gradient sequence and optimized material distribution curve of all sub-regions are integrated to form an initial parameter set.

3. The energy-saving and environmentally friendly green building design method as described in claim 2, characterized in that, The steps for developing energy consumption control models and forming environmental protection control models include: The optimized energy consumption intensity gradient sequence is input into the preset regression model. After multiple training iterations, the hidden layer weights are updated to generate the intensity compensation function in the energy consumption regulation model. The optimized material distribution curve is input into the clustering model, and the cluster centers are optimized through the backpropagation algorithm to generate the distribution correction mapping table in the environmental protection regulation model. After the regression model and clustering model converge, the feature vectors of the intensity compensation function and the distribution correction mapping table are extracted respectively. The matching degree between the feature vector and real-time meteorological parameters is calculated to verify the applicability of the energy consumption control model and the environmental protection control model. When the matching degree is lower than the set standard, the number of training iterations of the regression model and clustering model is readjusted until the feature vector meets the applicable conditions.

4. The energy-saving and environmentally friendly green building design method as described in claim 3, characterized in that, The specific steps for dynamically selecting a model and outputting adjustment instructions are as follows: Continuously track the changing trends of structural deviations and the oscillation range of ecological indicators during the design process; When the structural deviation trend exceeds the first threshold and the ecological index oscillation is within a stable range, the intensity compensation function in the energy consumption regulation model is activated. Based on the gradient adjustment rule in the intensity compensation function, an intensity correction command for the energy consumption device is generated. When the ecological indicators oscillate beyond the second threshold and the structural deviation trend is within a stable range, the distribution correction mapping table in the environmental protection regulation model is activated. Based on the distribution rules in the distribution correction mapping table, generate the distribution correction instruction for the material selection device; If both the structural deviation trend and the ecological indicator oscillation exceed the limit, the intensity correction command generated by the energy consumption control model will be executed first, and the command of the environmental protection control model will be suspended until the energy consumption adjustment is completed.

5. The energy-saving and environmentally friendly green building design method as described in claim 4, characterized in that, The steps for injecting and correcting the adjustment instructions into the execution process include: According to the gradient value in the intensity correction instruction, adjust the real-time intensity of each area in the energy consumption device step by step; After each adjustment, feedback data of the building structure is collected and the deviation is compared with the predicted value of the strength compensation function; If the deviation decreases, continue the current adjustment direction until the target structure is reached; If the deviation increases, the adjustment direction is reversed and the parameter update of the intensity compensation function is triggered again; Based on the distribution parameters in the distribution correction instruction, the selection speed of the material selection device in different sections is dynamically adjusted. During the distribution adjustment period, ecological indicators are monitored in real time using environmental sensors, and the coefficients in the distribution correction mapping table are updated based on the monitoring results.

6. The energy-saving and environmentally friendly green building design method as described in claim 1, characterized in that, The method further includes a feedback calibration phase after the adjustment instruction is executed, the feedback calibration phase including: Collect final structural data and ecological indicator maps after the completion of green building projects; The final structural data is compared with the prediction range of the energy consumption control model to generate an energy consumption model error signal; By performing overlap analysis between the ecological indicator map and the expected template of the environmental protection regulation model, an error signal for the environmental protection model is generated. Adjust the gradient compensation rule in the energy consumption control model based on the systematic error component in the energy consumption model error signal; Based on the random error component in the environmental protection model error signal, optimize the material distribution correction coefficient in the environmental protection control model; The updated gradient compensation rules and material distribution correction coefficients are synchronized to the historical database of the initial parameter set.

7. The energy-saving and environmentally friendly green building design method as described in claim 6, characterized in that, The process of adjusting the gradient compensation rules and material distribution correction coefficients specifically includes: Identify the stable deviation component in the energy consumption model error signal and calculate the compensation amount using the moving average method; Adjust the baseline gradient in the intensity compensation function according to the compensation amount; Identify high-frequency fluctuation components in the error signal of the environmental protection model and extract effective correction quantities through filtering techniques; Adjust the distribution weights in the distribution correction mapping table according to the effective correction amount; Replace the original parameters with the updated intensity compensation function and distribution correction mapping table.

8. The energy-saving and environmentally friendly green building design method as described in claim 1, characterized in that, The method also includes a preprocessing phase before the project begins, wherein the preprocessing phase is performed as follows: The climate type code of the target environment is parsed, the climate symmetry identifier and meteorological element code are extracted, and a climate feature vector is generated. The climate feature vector is input into a pre-stored environmental characteristic database for multi-dimensional similarity search, and a candidate template set with a similarity to the current climate type higher than a threshold is selected. For each candidate template, extract the average structural deviation and ecological indicator compliance rate from its historical records, and calculate the comprehensive effectiveness score; The candidate templates are sorted according to the scores, and the template with the highest score is selected as the benchmark energy intensity template. Extract the set of material distribution curves associated with the benchmark energy intensity template from the database; For each curve, the consistency between distribution and intensity gradient is checked, and abnormal curves with abrupt changes or conflicts are removed to form an optimized set of material distribution curves. Based on the historical stability index of the optimized curve set, the curve with the smallest fluctuation is selected as the benchmark material distribution curve. The baseline energy intensity template and the baseline material distribution curve are time-aligned to generate a baseline configuration for the initial parameter set.

9. The energy-saving and environmentally friendly green building design method as described in claim 8, characterized in that, The steps for verifying the compatibility of material distribution curves include: Extract individual curve data from the curve set and obtain the intensity gradient sequence corresponding to time in the benchmark energy intensity template; Based on the time points of the intensity gradient sequence, collaborative timestamps are marked on the curve to generate a marked distribution curve; Traverse the marked curves and detect whether the slope of the distribution change within adjacent time intervals exceeds the limit to identify abnormal intervals; When an abnormal interval is detected, the gradient value at the corresponding time point in the intensity gradient sequence is traced back to determine whether the gradient fluctuation direction conflicts with the direction of the sudden change in distribution. If the conflict intensity exceeds the limit, it is marked as a conflict area, and its starting position and duration are recorded. Based on the location and duration of the conflict area, a correction interval is defined on the curve, and a smooth replacement segment is generated based on historical data. Insert the replacement segment into the correction interval to generate the optimized curve, and delete the original abnormal data points; After completing the replacement operation on all curves, perform an integrity check and remove any uncorrected residual curves. The obtained curves are merged into an optimized material distribution curve set.

10. The energy-saving and environmentally friendly green building design method as described in claim 1, characterized in that, The method further includes a real-time optimization iteration phase after the adjustment instructions are generated, the real-time optimization iteration phase including the execution of: Continuously collect new environmental monitoring data during project implementation and update the initial parameter set; Based on the updated parameter set, the applicability of the energy consumption control model and the environmental protection control model are re-verified. When the verification fails, the model retraining process is initiated to regenerate the intensity compensation function and distribution correction mapping table using the latest data. The retrained model parameters are integrated into the tuning instructions to form a closed-loop optimization mechanism.

Images