Digital intelligent green building design evaluation method

Abstract

The invention relates to the technical field of green building design, and discloses a digital intelligent green building design evaluation method. The method comprises the steps of collecting a multi-dimensional design parameter data set of a target building, wherein the multi-dimensional design parameter data set comprises spatial layout, material attributes, energy consumption and environment interference event parameters; energy consumption characteristic parameters, structural characteristic parameters and interference characteristic parameters corresponding to environment interference events are separated from the data set; establishing dynamic correlation characteristics of energy consumption and structural characteristic parameters, and extracting cooperative adjustment characteristics of environmental interference events on the two parameters; generating a comprehensive evaluation feature set representing the sustainability difference of the building based on the features, and calculating environmental performance score distribution; and according to the distribution optimization multi-dimensional design parameter data set, an optimization energy consumption and structural characteristic curve is separated. According to the method, through multi-dimensional parameter analysis and dynamic association mining, the comprehensiveness and accuracy of evaluation are improved, and a direction is indicated for design optimization.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of green building design, in particular to a digital intelligent green building design evaluation method. BACKGROUND

[0002] With the acceleration of global urbanization, the construction industry has increasingly paid attention to energy consumption and environmental impact, and green building design has gradually become the core direction of industry development. Traditional green building design evaluation methods mostly rely on static index systems, qualitatively or semi-quantitatively evaluate building design schemes through preset scoring standards, and are difficult to fully reflect the dynamic performance of buildings in the whole life cycle. In practical application, traditional methods often focus on single-dimensional parameter analysis, such as only focusing on energy consumption indicators or material environmental protection properties, ignoring the correlation between parameters. For example, the spatial layout of a building not only affects lighting and ventilation, but also indirectly changes the energy consumption mode, and traditional evaluation methods are difficult to capture the dynamic correlation between multi-dimensional parameters. Buildings are affected by various environmental disturbance events during operation, such as extreme weather, changes in surrounding traffic flow, etc., which have a synergistic effect on the energy consumption efficiency and structural stability of the building. Traditional evaluation methods usually analyze environmental disturbances as independent factors, and cannot accurately extract their comprehensive adjustment effect on energy consumption characteristics and structural characteristics, resulting in deviations between evaluation results and actual operation. Traditional evaluation methods mostly use fixed scoring models, which are difficult to dynamically generate differentiated evaluation features according to the specific parameters of the building, resulting in a lack of pertinence when guiding design optimization. When the design scheme needs to be adjusted, it often needs to be re-evaluated throughout the process, which is inefficient and difficult to meet the needs of modern green building design for precision and efficiency. SUMMARY

[0003] The purpose of the present application is to provide a digital intelligent green building design evaluation method to solve the problems raised in the background art.

[0004] To achieve the above purpose, the present application provides a digital intelligent green building design evaluation method, which comprises: Collecting multi-dimensional design parameter data sets of the target building, the multi-dimensional design parameter data sets including spatial layout parameters, material attribute parameters, energy consumption parameters and environmental disturbance event parameters; Separating energy consumption characteristic parameters, structural characteristic parameters and disturbance characteristic parameters corresponding to environmental disturbance events from the multi-dimensional design parameter data sets; Establishing dynamic correlation characteristics between the energy consumption characteristic parameters and the structural characteristic parameters, and extracting the synergistic adjustment characteristics of the environmental disturbance events on the energy consumption characteristic parameters and the structural characteristic parameters; generate a set of comprehensive evaluation features representing the differences in building sustainability based on the dynamic correlation features and the synergistic regulation features; and calculate an environmental performance score distribution of the building design scheme according to the set of comprehensive evaluation features; perform data optimization processing on the multi-dimensional design parameter dataset according to the environmental performance score distribution, and separate out an optimized energy consumption feature curve and an optimized structural feature curve of the target building.

[0005] Preferably, the separation of the energy consumption feature parameters, the structural feature parameters, and the disturbance feature parameters corresponding to the environmental disturbance events from the multi-dimensional design parameter dataset comprises the following steps: perform spatial domain division on the multi-dimensional design parameter dataset using a dynamic clustering algorithm, construct an energy consumption feature main channel, and retain the periodic energy consumption fluctuation components within the range of 0.05-1.2 Hz through frequency band screening; establish a structural feature auxiliary channel based on adaptive weighted fusion, and use a Gaussian kernel function to perform multi-scale decoupling on the structural response data within the frequency band of 1.2-5 Hz; evaluate the independence indicators of the energy consumption feature main channel and the structural feature auxiliary channel through mutual entropy measurement, dynamically adjust the fusion weights between the channels to realize parameter separation, and output the energy consumption feature parameters and the structural feature parameters.

[0006] Preferably, the extraction of the disturbance feature parameters corresponding to the environmental disturbance events from the multi-dimensional design parameter dataset comprises the following steps: perform mutation gradient detection on the time series frames of the multi-dimensional design parameter dataset, locate the candidate time intervals of the environmental disturbance events, and generate event markers; perform non-stationary feature decomposition on the marker data corresponding to the event markers, and extract the feature components within the frequency band of 1-10 Hz reflecting external environmental disturbances; distinguish local environmental disturbances from systematic noise based on a spatiotemporal correlation matrix, fuse the frequency band feature components to generate a disturbance intensity time series curve, and the disturbance intensity time series curve is the disturbance feature parameter.

[0007] Preferably, the mutation gradient detection on the time series frames of the multi-dimensional design parameter dataset, the location of the candidate time intervals of the environmental disturbance events, and the generation of the event markers comprise the following steps: calculate the gradient energy operator difference sequence of adjacent time series frames through a sliding analysis window; generate a dynamic decision threshold according to the real-time data fluctuation characteristics; determine the starting time point and the duration range of the environmental disturbance events based on the dynamic decision threshold; Locate a candidate time interval using the starting time point and the duration range, and generate an environmental interference event label.

[0008] Preferably, the spatial-temporal correlation matrix is used to distinguish local environmental interference from systematic noise, and the frequency band feature components are fused to generate an interference intensity time series curve, including the following steps: According to the phase changes of multi-source data, a three-dimensional spatial-temporal correlation graph is constructed, and the covariance distance of each spatial node and the reference trajectory is calculated. When the cumulative covariance distance of the continuous spatial nodes exceeds the dynamic noise threshold, it is determined as a local environmental interference event and the spatial positioning coordinate set is output. The instantaneous energy distribution of the frequency band feature component is spatio-temporally aligned with the spatial positioning coordinate set to generate a time-continuous interference intensity time series curve.

[0009] Preferably, the dynamic correlation feature between the energy consumption characteristic parameter and the structure characteristic parameter is established, including the following steps: The structure response coefficient of variation and the energy intensity-structure deformation correlation index in the energy consumption fluctuation period are calculated to generate a dynamic correlation feature. An environmental interference-energy consumption coupling criterion is established to analyze the stability index and structure recovery rate of the energy consumption characteristic parameter before and after the environmental interference event, and a cooperative regulation feature is generated.

[0010] Preferably, before the comprehensive evaluation feature set representing the differences in building sustainability is generated, the following steps are further included: It is judged whether the duration range of the environmental interference event exceeds the preset duration threshold, and whether there is a multi-building reflection cross section overlap. The comprehensive evaluation feature set representing the differences in building sustainability is generated, including the following steps: If the duration range of the environmental interference event does not exceed the preset duration threshold and there is no multi-building reflection cross section overlap, an initial evaluation feature set is generated according to the dynamic correlation feature and the cooperative regulation feature, and the initial evaluation feature set is used as the comprehensive evaluation feature set.

[0011] Preferably, the method further includes the following steps: If the duration range of the environmental interference event exceeds the preset duration threshold, or there is a multi-building reflection cross section overlap, a dynamic clustering algorithm is used to perform secondary decoupling processing on the energy consumption characteristic parameter and the structure characteristic parameter. Based on the three-dimensional spatial-temporal correlation graph, the reflection cross section overlap area corresponding to the systematic noise is excluded, and the processed energy consumption characteristic parameter and the processed structure characteristic parameter are re-extracted. According to the processed energy consumption feature parameter, a target dynamic correlation feature is generated, and according to the processed structure feature parameter, a target synergistic regulation feature is generated; the target dynamic correlation feature and the target synergistic regulation feature are used to update the initial evaluation feature set.

[0012] Preferably, after the secondary decoupling processing of the energy consumption feature parameter and the structure feature parameter by using the dynamic clustering algorithm, the following steps are further included: The energy distribution convergence of the data after the secondary decoupling processing is verified; If the verification is passed, the initial evaluation feature set is corrected based on the target dynamic correlation feature and the target synergistic regulation feature, an updated evaluation feature set is generated, and the updated evaluation feature set is taken as a comprehensive evaluation feature set.

[0013] Preferably, the data optimization processing of the multi-dimensional design parameter data set according to the environmental performance score distribution includes the following steps: The multi-dimensional design parameter data set is subjected to weight filtering based on the environmental performance score distribution, and the energy consumption and structure interference components of non-target buildings are suppressed to generate an optimized data set; The optimized data set is subjected to a space-time feature separation algorithm to extract an optimized energy consumption feature curve and an optimized structure feature curve of the target building.

[0014] Compared with the prior art, the present application has the following beneficial effects: By collecting the multi-dimensional design parameter data set, key factors such as spatial layout, material properties, energy consumption and environmental disturbance events are covered, and the building design is fully described. Compared with the evaluation method that only focuses on a single parameter, this multi-dimensional data collection can more completely reflect the overall characteristics of building design and provide rich basic information for subsequent evaluation. In the parameter processing link, the energy consumption feature parameter, the structure feature parameter and the interference feature parameter are separated, and the dynamic correlation feature of the energy consumption and structure feature parameters is further established, which breaks through the static cognition of the relationship between parameters in the traditional evaluation. By capturing the dynamic correlation between the two generated by the design parameters, the internal relationship between the energy flow and the structure performance in building design can be more accurately revealed, and the evaluation result is more in line with the actual operation logic of the building. For environmental disturbance events, the method extracts the synergistic regulation feature of the energy consumption and structure features, avoiding the limitation of isolated analysis of environmental factors. The influence of environmental disturbance events on buildings is often comprehensive, which may change the energy consumption mode and affect the stress state of the structure. By analyzing the synergistic regulation effect, the actual performance of the building under complex environmental conditions can be more accurately evaluated, and the authenticity and reliability of the evaluation are improved. The comprehensive evaluation feature set generated based on the dynamic correlation feature and the synergistic regulation feature can represent the difference of building sustainability, and provides differentiated evaluation basis for the environmental performance score distribution of the design scheme. This differentiated evaluation method breaks free from the shackles of the fixed scoring model, and can generate targeted evaluation results according to the specific features of different building designs, so that the design personnel can clearly identify the advantages and disadvantages of different schemes in terms of sustainability. In the design optimization stage, the multidimensional design parameter data set is optimized according to the environmental performance score distribution, and the optimized energy consumption feature curve and the optimized structure feature curve are separated, which provides a clear direction for the adjustment of the design scheme. Through the optimization curve directly pointing to the energy consumption and structure features, the design personnel can adjust the relevant design parameters, reduce the blindness in the optimization process, make the design optimization more targeted, and help to meet the requirements of green building while taking into account the functionality and economy of the building.

[0015] The whole evaluation process adopts digital intelligent processing method, realizes the intelligentization of the whole process from data acquisition, feature extraction to evaluation optimization, reduces the subjective bias caused by manual intervention, and improves the evaluation efficiency. When the design scheme needs to be adjusted, new evaluation results can be quickly generated based on the existing parameter correlation and evaluation model, which shortens the optimization cycle and meets the demand of rapid iteration in modern green building design. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The timing diagram of the digital intelligent green building design evaluation method described in the present application; Figure 2 The flowchart for separating energy consumption and structure feature parameters; Figure 3 The flowchart for generating environmental disturbance event markers; Figure 4 The flowchart for generating dynamic correlation and synergistic regulation features; Figure 5 The flowchart for secondary decoupling and feature updating. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0018] Please refer to Figure 1 The present application provides a digital intelligent green building design evaluation method, which comprises: A multi-dimensional design parameter dataset of the target building is collected, which includes spatial layout parameters, material attribute parameters, energy consumption parameters, and environmental disturbance event parameters. Energy consumption feature parameters, structure feature parameters, and disturbance feature parameters corresponding to environmental disturbance events are separated from the multi-dimensional design parameter dataset. A dynamic correlation feature between the energy consumption feature parameters and the structure feature parameters is established, and a synergistic regulation feature of the environmental disturbance events on the energy consumption feature parameters and the structure feature parameters is extracted. Based on the dynamic correlation feature and the synergistic regulation feature, a comprehensive evaluation feature set representing the differences in building sustainability is generated. The environmental performance score distribution of the building design scheme is calculated according to the comprehensive evaluation feature set. According to the environmental performance score distribution, data optimization processing is performed on the multi-dimensional design parameter dataset, and the optimized energy consumption feature curve and the optimized structure feature curve of the target building are separated.

[0019] Embodiment 1: see Figure 2 Separating the energy consumption feature parameters, the structure feature parameters, and the disturbance feature parameters corresponding to the environmental disturbance events from the multi-dimensional design parameter dataset involves a series of data processing and analysis steps. The process first divides the spatial domain of the multi-dimensional design parameter dataset using a dynamic clustering algorithm to construct an energy consumption feature main channel. The dynamic clustering algorithm automatically groups data points based on their spatial distribution characteristics, classifying parameters with similar features into the same cluster to form the energy consumption feature main channel. This channel is mainly used to extract periodic fluctuation components related to building energy consumption, and effective signals within the 0.05-1.2 Hz range are retained through frequency band filtering. Frequency band filtering is achieved using a digital filter to exclude high-frequency noise and low-frequency drift components, ensuring the stability of the energy consumption feature.

[0020] At the same time as constructing the energy consumption feature main channel, a structure feature auxiliary channel based on adaptive weighted fusion is established. This channel is used to process structure response data in the 1.2-5 Hz frequency band, and a Gaussian kernel function is used for multi-scale decoupling. The scale parameter of the Gaussian kernel function is dynamically adjusted according to the data fluctuation characteristics to adapt to the changes in structure features in different spatial regions. An adaptive weighting mechanism is used to fuse the decoupled data, and the weight distribution is based on local variance calculation to ensure the accuracy of structure feature extraction. The independence index between the energy consumption feature main channel and the structure feature auxiliary channel is evaluated by mutual entropy measurement, and the fusion weight between the channels is dynamically adjusted to finally output the separated energy consumption feature parameters and structure feature parameters. Mutual entropy measurement calculates the degree of information overlap between the two channels. If the independence is insufficient, the fusion weight is adjusted to enhance the feature separation effect.

[0021] The interference feature parameter extraction of the environmental interference event is realized through time series analysis and non-stationary signal processing. First, the mutation gradient of the time series frame of the multi-dimensional design parameter data set is detected, and the candidate time interval of the environmental interference event is located. The mutation gradient detection calculates the gradient energy change of adjacent frames based on the sliding analysis window, and when the gradient difference exceeds the dynamic judgment threshold, it is marked as a potential interference event. The dynamic judgment threshold is generated according to the real-time data fluctuation characteristics, avoiding false judgment or missed detection caused by fixed threshold. The candidate time interval is determined by the starting time point and the duration range, and the corresponding event marker is generated.

[0022] Non-stationary feature decomposition is performed on the marked data to extract environmental interference feature components in the 1-10Hz frequency band. The non-stationary feature decomposition uses an adaptive signal decomposition method to decompose the marked data into multiple intrinsic mode functions, and selects the frequency band components reflecting external environmental disturbance. Based on the space-time correlation matrix, local environmental interference and systematic noise are further distinguished. The space-time correlation matrix considers the phase change relationship of multi-source data, and calculates the covariance distance of each spatial node and the reference trajectory. If the cumulative covariance distance of consecutive spatial nodes exceeds the dynamic noise threshold, it is determined as a local environmental interference event, and the spatial positioning coordinate set is output. The dynamic noise threshold is adaptively set by the current system noise level to ensure the robustness of interference detection.

[0023] The interference feature parameters are generated in the form of interference intensity time curve. This curve is realized by fusing the 1-10Hz frequency band feature components and the spatial positioning coordinate set, and the space-time alignment ensures the continuity and consistency of the interference intensity. The interference intensity time curve is used as the quantitative representation of the environmental interference event for subsequent dynamic correlation analysis and comprehensive evaluation.

[0024] The construction of the energy consumption feature main channel depends on the spatial domain division ability of the dynamic clustering algorithm. The algorithm calculates the similarity according to the multi-dimensional attributes of the data points, and automatically divides the regions with consistent energy consumption characteristics. The frequency band selection uses a digital band-pass filter to retain the periodic energy consumption fluctuation components in the 0.05-1.2Hz range, which covers the main variation mode of building energy consumption. The Gaussian kernel function of the structural feature auxiliary channel decouples the structural response data of different spatial scales, ensuring the balanced extraction of local and global features. The adaptive weighted fusion mechanism adjusts the weight according to the local characteristics of the data, avoiding the dominance of a single scale feature in the analysis results.

[0025] The detection of environmental interference events uses a gradient energy operator to calculate the change intensity of adjacent time sequence frames, dynamically determines the threshold value to adapt to the data fluctuation range, and reduces the false positive rate. The non-stationary feature decomposition extracts the 1-10Hz frequency band component to reflect the typical frequency characteristics of environmental interference, the construction of the space-time correlation matrix integrates the spatial and temporal dimension information, and the covariance distance calculation distinguishes real interference from system noise. The generation of the interference intensity time sequence curve combines the frequency band energy distribution and spatial positioning information to quantify the influence degree of the interference event.

[0026] The entire implementation process involves dynamic clustering, frequency band filtering, multi-scale decoupling, mutual entropy measurement, gradient detection, non-stationary decomposition, and space-time correlation analysis, etc. Various data processing techniques ensure the effective separation of energy consumption feature parameters, structural feature parameters, and interference feature parameters. The separated feature parameters are used for subsequent dynamic correlation feature construction and comprehensive evaluation, supporting intelligent analysis and optimization decision of green building design.

[0027] In the construction process of the energy consumption feature main channel, the dynamic clustering algorithm uses an iterative optimization method to adjust the cluster center until the data point assignment is stable. The bandpass filter design of frequency band screening considers the transition band attenuation characteristics to avoid signal distortion at the frequency band boundary. The scale selection of the Gaussian kernel function of the structural feature auxiliary channel is based on the local statistical characteristics of the data to ensure reasonable expression of multi-scale features. The weight update of adaptive weighted fusion is based on local variance changes to dynamically balance the contribution of different scale features. The calculation of mutual entropy measurement uses a sliding window method to monitor the channel independence in real time and adjust the fusion strategy.

[0028] The detection of environmental interference events relies on the difference calculation of the gradient energy operator, and the dynamic determination of the threshold value is adaptively adjusted according to the statistical characteristics of the historical data to reduce the misjudgment caused by environmental fluctuations. The non-stationary feature decomposition uses adaptive basis function expansion to ensure accurate extraction of interference features. The construction of the space-time correlation matrix integrates multi-source sensor data, and the covariance distance calculation considers the spatial topological relationship to improve the positioning accuracy of the interference. The generation of the interference intensity time sequence curve uses an interpolation algorithm to smooth the energy distribution, ensuring time continuity.

[0029] This implementation realizes the efficient separation of energy consumption features, structural features, and interference features through multi-step data processing and analysis, providing a reliable data foundation for intelligent evaluation of green building design. The separated feature parameters are further used for dynamic correlation analysis and environmental performance score calculation, supporting sustainable optimization decision of buildings. The entire process uses an automated processing method to reduce human intervention, improve analysis efficiency and accuracy.

[0030] Example 2: see Figure 3The operation of mutation gradient detection on the time series frame of multi-dimensional design parameter dataset involves specific data processing mechanism. The time series is traversed through a fixed-size sliding analysis window, and the gradient energy operator difference between adjacent frames is calculated in each window. The gradient energy operator is constructed based on the second-order differential feature, reflecting the degree of parameter change. The step size of the sliding window matches the data acquisition frequency, ensuring continuous coverage of all time nodes. The difference sequence obtained by calculation forms the energy change trajectory in the time dimension, providing a quantitative basis for subsequent event determination.

[0031] The generation mechanism of dynamic determination threshold depends on the statistical characteristics of real-time data stream. The standard deviation of the current parameter is calculated using the historical data window, and an adaptive threshold model is established combined with the long-term fluctuation baseline. This model introduces a decay factor to adjust the weight of historical data, so that the threshold can track the dynamic changes of data distribution. When the peak value of the difference sequence exceeds the dynamically calculated dynamic determination threshold, the event recognition process is triggered. The event starting time point is defined as the moment when the difference first exceeds the threshold, and the duration range is determined by recording the length of the continuous super-threshold period. The time interval boundaries are marked with precise time stamps, and finally the environmental disturbance event markers containing the start and end time indexes are generated.

[0032] In the spatial dimension, the interference feature extraction relies on the construction of three-dimensional spatiotemporal correlation graph. The graph integrates multi-source data collected by distributed sensor networks to construct a three-dimensional matrix indexed by spatial coordinates and time stamps. The reference trajectory is defined as the standard variation pattern of the parameter under the non-interference state, which is modeled through historical data statistical analysis. For each spatial node, the covariance distance between its parameter variation and the reference trajectory is calculated, which reflects the degree of deviation of the current node behavior from the standard pattern. The distance calculation considers the correlation between parameters and uses multivariate statistical methods to handle cross-influence.

[0033] The identification of local environmental disturbance uses spatial continuity criterion. When the cumulative covariance distance value of a cluster composed of adjacent spatial nodes exceeds the dynamic noise threshold, it is determined that there is local disturbance in this region. Spatial proximity is defined by a pre-set connected distance, and node clusters are divided according to the rules of graph theory connected components. The dynamic noise threshold is updated according to the real-time noise level of the system, and the noise level is evaluated in real time through statistical analysis of background parameter changes. The identified interference events output contain a set of spatial positioning coordinates, which describe the spatial range of the interference region in the Cartesian coordinate system.

[0034] The generation of the interference intensity time series curve requires the fusion of the results of the analysis in the time dimension and the space dimension. The marker data corresponding to the event markers are subjected to non-stationary feature decomposition, and the feature components in the 1-10 Hz frequency band are extracted. These components represent the spectral characteristics of environmental interference, and the intensity variation in the time dimension is obtained by calculating the instantaneous energy distribution. The spatial positioning coordinate set provides the physical location information of the interference event. The time-space alignment operation establishes the mapping relationship between the time index and the spatial coordinates, so that the energy values of the frequency band feature components can be accurately matched to the interference area. The alignment process uses time synchronization interpolation processing to solve the timestamp difference problem of different acquisition devices.

[0035] The final generated interference intensity time series curve is complete time series data. Each point on the curve contains a time label and an intensity value, and the intensity value comprehensively reflects the interference level of the spatial area at a specific time. This curve is used as a quantitative representation for subsequent analysis, recording the dynamic change process of environmental interference events. The entire implementation process uses an automated processing flow, from data acquisition to curve generation without human intervention, maintaining processing efficiency and analysis consistency.

[0036] The sliding analysis window size design of the mutation gradient detection link considers the physical characteristics of parameter changes. The window length covers the typical transient process duration, and too short will cause noise sensitivity, and too long will reduce the time resolution. The gradient energy operator calculation optimizes the calculation efficiency, using discrete difference approximation to differential operation, balancing accuracy and calculation load. The decay factor in the dynamic threshold model is set through multiple parameter optimization experiments, so that the threshold response speed matches the environmental change speed.

[0037] The data storage structure of the three-dimensional space-time correlation graph uses sparse matrix technology, only recording the position and data value of non-zero elements. This design effectively reduces memory occupancy, supporting data processing of large-scale building groups. The covariance distance calculation uses an iterative algorithm to update gradually, avoiding resource consumption for full calculation each time. The connected distance threshold of spatial proximity is set according to the physical connection characteristics of the building structure, such as the structure seam spacing or functional partition boundary.

[0038] The space-time alignment processing of interference events includes coordinate transformation operations. When the sensor layout coordinate system and the building physical coordinate system are inconsistent, affine transformation is used to complete the coordinate system conversion. The instantaneous energy calculation of the frequency band feature components uses windowed Fourier transform, with the window length matching the event duration. Time synchronization interpolation is implemented based on the Lagrange interpolation algorithm, ensuring the consistency of multi-source data in the time dimension.

[0039] The data structure of the interference intensity time series curve includes a timestamp array and an intensity value array, which maintain an index correspondence. The intensity value normalization process eliminates the dimensional influence, and the data range is controlled between 0 and 1. The curve generation module outputs a standard format data file, which can be seamlessly connected with other system components. The abnormal value filtering mechanism in the data processing chain uses a sliding median filter to remove sudden noise points.

[0040] This embodiment establishes a complete environmental disturbance event detection and quantification process, covering the conversion process from raw data to feature curves. The time analysis module identifies the disturbance time interval, the spatial analysis module locates the disturbance spatial range, and the fusion module generates complete spatio-temporal features. The parameters of each module are preset based on building data analysis experience, which meets the requirements of typical green building evaluation scenarios. The processing flow adopts a pipeline architecture, and the output of each stage can be independently verified, supporting modular debugging and optimization. The entire system maintains stable memory usage and processing delay when processing continuous data streams, meeting real-time monitoring requirements.

[0041] Example 3: Refer to Figure 4 The dynamic correlation feature between energy consumption characteristic parameters and structural characteristic parameters includes a quantitative analysis process. The coefficient of variation of structural response in the energy consumption fluctuation period is calculated, which is obtained by dividing the standard deviation of the structural characteristic parameter in a specific time period by its arithmetic mean. The energy consumption fluctuation period is identified using spectral analysis method, and the period length corresponding to the dominant frequency is taken as the calculation interval. The structural response data comes from a distributed sensor network, and sample points are continuously collected within the identified period. The variation coefficient calculation result reflects the relative change of the structural parameter, and the numerical value indicates the sensitivity of the structural response to external energy consumption changes. At the same time, the correlation index between energy intensity and structural deformation is calculated, and the energy intensity is quantified by the instantaneous power consumption value per unit area of the building, and the structural deformation is obtained by millimeter-level precision measurement of displacement sensors. The correlation analysis uses Pearson linear correlation coefficient The formula is defined as: Wherein: represents the energy consumption intensity value at the i-th time point, is the structural deformation value at the corresponding time point, and represent the average value of the time series, respectively, and n is the total number of data sampling points in the current fluctuation period. The calculation result forms a dynamic correlation feature vector, which includes a two-component combination of the variation coefficient and the correlation coefficient.

[0042] The environmental disturbance event analysis uses the environmental disturbance-energy consumption coupling criterion. The criterion sets the disturbance intensity threshold When the mean deviation of the disturbance characteristic parameters exceeds The analysis mechanism is triggered when the interference occurs. In the time window from 300 seconds before to 600 seconds after the interference event, the energy consumption characteristic parameter data segments are intercepted respectively. The stability index calculation is completed by the ratio of the variance of the data segments before and after the event. If the variance decreases after the event, the index is greater than 1, and if the variance increases, it is less than 1. The structural recovery rate analysis focuses on the key time period of 0-200 seconds after the interference occurs. The least squares method is used to fit the attenuation curve of the structural characteristic parameters. The absolute value of the linear fitting slope k defines the recovery rate value. The larger the slope, the stronger the recovery ability. The collaborative adjustment feature is composed of the product of the stability index and the recovery rate. This numerical combination reflects the adaptive adjustment ability of the building system under environmental interference.

[0043] Implement pre-set condition determination before executing building sustainability gap analysis. Pre-set time threshold Set to a fixed value of 600 seconds, by comparing the duration of environmental interference events and The overlap detection of multiple building reflection sections uses pulse laser ranging data. When the distance between the adjacent building reflection surfaces is Less than the safety threshold The detection algorithm uses the bounding box collision detection principle to scan the minimum bounding box of the building surface in the spatial point cloud data.

[0044] The operation of generating comprehensive evaluation feature set is divided into path selection. If there is no overlap of multiple buildings, the initial evaluation feature set is constructed directly based on the dynamic correlation feature vector and the coordinated adjustment feature value. The set consists of three-dimensional vectors, the first dimension of which is the coefficient of variation of the structural response. , the second dimension is the energy consumption-deformation correlation coefficient , the third dimension is the collaborative regulation eigenvalue Before the vector is formed, normalization is performed, and the values ​​of each dimension are normalized to the interval [0, 1]. The comprehensive evaluation feature set maintains the same structure as the initial set, and the positions of the vector elements remain unchanged.

[0045] The above process involves controlling the accuracy of spectral analysis when identifying energy consumption fluctuation periods. Dominant frequencies are identified using a Hanning windowed fast Fourier transform, with a spectral resolution of 0.001 Hz, ensuring a period measurement error of less than 0.5 seconds. The coefficient of variation of the structural response is calculated using rolling window sampling, with the window size dynamically adjusted with the fluctuation period, maintaining a constant number of sampling points at 200. Data synchronization calibration is performed before correlation coefficient calculation. Clock offsets between the energy consumption and deformation data channels are automatically corrected using a cross-correlation function, ensuring time delay compensation accurate to the millisecond level.

[0046] The stability index calculation adopts piecewise variance estimation. The pre-event section selects 150 evenly distributed sampling points within 300 seconds before the disturbance occurs. The post-event section selects 450 sampling points within 200-800 seconds after the disturbance ends. The difference in data segment length is balanced by the degree of freedom correction factor to balance the variance calculation results. The structure recovery rate fitting adopts weighted least squares method. Data points close to the event occurrence time are assigned higher weights to reflect the dynamic characteristics in the early recovery period. The trigger control of the preset time threshold is realized by a hardware timer. The value writing system configuration register supports dynamic updating.

[0047] The multi-building reflection cross-section overlap detection sets a double verification mechanism. After generating the spatial point cloud by pulse laser scanning, the first detection is to detect the angle between the normal vectors of the building surfaces When the angle is less than 5 degrees, and the distance is less than 2 meters, the overlap confirmation process is entered. The second detection uses infrared thermal imaging to assist verification. If the temperature field difference between the target building and the adjacent building is below the set threshold, it is determined to be a real overlap. The safety threshold is set to 2.5 meters for commercial buildings and 1.8 meters for residential buildings according to the building category configuration.

[0048] The standardization method for constructing the comprehensive evaluation feature set adopts the maximum and minimum value scaling. Each element in the dynamic correlation feature vector is independently standardized: The value is divided by the preset upper limit coefficient 3.0, the original value range [-1, 1] is mapped to the [0, 1] interval, and the value is processed using logarithmic scaling to compensate for the order of magnitude difference. The processed three-dimensional vector is directly stored in the evaluation result buffer, waiting for subsequent environmental performance score calculation calls. The entire process sets an integrity check step to verify the time window alignment state and data segment validity flag, preventing calculation errors caused by abnormal input. The vector generation module establishes a streaming interface with the upstream data pipeline to support incremental calculation under real-time data update conditions.

[0049] Example 4: see Figure 5 Taking the environmental disturbance event analysis of a certain commercial complex building group as an example, when the system detects that the duration exceeds the preset threshold or there is a multi-building reflection cross-section overlap, the secondary decoupling processing flow is triggered. The building group includes a main building (Building A) and two auxiliary buildings (Buildings B and C), which are arranged in a triangular shape with close spacing. A certain environmental disturbance event caused by strong wind lasted for 720 seconds, exceeding the preset threshold of 600 seconds. At the same time, laser scanning detected that Buildings B and C had a reflection cross-section overlap region between the 15th and 20th floors.

[0050] The system first performs dynamic clustering analysis on the original collected energy consumption characteristic parameters and structural characteristic parameters. The improved DBSCAN algorithm is used to process the three-dimensional parameter space, and the core parameters are set as follows: neighborhood radius ε = 0.45, minimum sample number MinPts = 12. The clustering process generates three main clusters, and their distribution characteristics are shown in Table 1.

[0051] Table 1: The distribution characteristics of the three main clusters are as follows.

[0052] Analysis found that cluster 2 and cluster 3 have obvious overlap in the parameter space, corresponding to the overlapping area of B and C buildings. The system starts three-dimensional spatiotemporal correlation map analysis, and constructs a correlation matrix containing 1.2 million voxels with a spatial resolution of 0.5 meters. By calculating the covariance distance of each voxel and the reference trajectory, it is found that the specific range of the overlapping area is between 62-68 meters in height, corresponding to the connecting corridor part of the 15-17 floors of the building.

[0053] The noise exclusion process uses a double-threshold filtering mechanism. First, set the covariance distance threshold D_cov = 1.8 to filter out systematic noise voxels; second, set the spatial continuity threshold S_con = 5 to require that valid voxels must form a continuous spatial block. The core area retained after processing includes three independent sub-areas: the west facade of Building A (area 420 m²), the east facade of Building B (area 380 m²), and the north facade of Building C (area 350 m²).

[0054] The re-extracted characteristic parameters show obvious changes. The fluctuation amplitude of the energy consumption characteristic parameters in the overlapping area is reduced by about 40% after processing, and the main periodic component is adjusted from 0.8 Hz to 0.6 Hz. The standard deviation of the structural characteristic parameters is reduced by 35%, and the peak frequency of the displacement response spectrum moves 0.3 Hz to the low frequency direction. These changes indicate that the secondary decoupling effectively separates the mutual interference between buildings.

[0055] The generation of target dynamic correlation features uses the updated parameter set. Within a typical energy consumption fluctuation period (1.67 seconds), the structural response coefficient of variation decreases from the initial value of 0.38 to 0.24, and the energy intensity-structure deformation correlation index increases from 0.62 to 0.71. The target cooperative regulation feature calculation shows that the stability index increases from 1.2 to 1.5 after the disturbance event, and the structure recovery rate increases by 22%. These numerical changes reflect that the parameters after decoupling processing can more accurately represent the inherent characteristics of single buildings.

[0056] The energy distribution convergence verification adopts the frequency band energy integration method. By comparing the power spectrum density curves before and after processing, the energy concentration degree is improved from 75% to 88% in the main frequency band of 0.1-5Hz, and the cross-band leakage is reduced by 60%. After verification, the system modifies the initial evaluation feature set according to the update weight of 0.7, generating a new three-dimensional feature vector (0.24, 0.71, 1.5).

[0057] This implementation process specially handles the parameter aliasing problem of building connection parts. The sensor data in the connecting corridor area contains the influence of both buildings, and the conventional analysis method is difficult to distinguish. Through the fine analysis of the space-time correlation graph, the system successfully separates the feature parameters of B and C buildings respectively. For example, in the wind speed disturbance response analysis, the originally mixed vibration mode is decomposed into the 0.7Hz bending mode of B building and the 0.9Hz torsional mode of C building.

[0058] The parameter updating mechanism adopts a gradual adjustment strategy. After each environmental disturbance event is processed, the system retains 20% of the historical parameter features and weights them with the newly extracted 80% feature values. This design avoids parameter mutation-induced evaluation fluctuations and maintains the continuity of building performance evaluation. At the same time, an outlier rejection rule is set, which triggers an artificial review process when the new feature value deviates from the historical mean value by more than three standard deviations.

[0059] The special processing for multi-building group scenarios also includes spatial reference system calibration. The system establishes a unified coordinate system with the geomagnetic north as the reference, and the sensor data of each building is first converted to this coordinate system before analysis. The coordinate conversion accuracy is controlled within ±0.3 degrees, ensuring the geometric accuracy of cross-building data correlation. In this case, this processing effectively eliminates the wind speed data deviation caused by the difference in building orientation.

[0060] The update frequency of the comprehensive evaluation feature set is set to an event-driven mode. The regular monitoring period keeps the initial set unchanged, and only when the secondary decoupling condition is triggered does the update process start. The update operation adopts a transaction processing mechanism to ensure the synchronous update of each dimension of the feature vector. The system maintains three versions of the feature set: the original version, the temporary processing version, and the confirmed version, supporting the backtracking verification of the analysis process.

[0061] The implementation case of this commercial complex shows that for building groups with strong mutual interference, the conventional single building analysis method will produce significant errors. Through the secondary decoupling processing of this example, the system can effectively identify and separate the cross-influence, extracting more accurate single building feature parameters. This processing is particularly suitable for complex scenarios such as urban high-density building groups and connected buildings, improving the relevance and reliability of green building evaluation. The entire processing flow is completed on standard server hardware, with an average processing time of 8 minutes and 30 seconds for a single event, meeting the timeliness requirements of practical engineering.

[0062] The operation of performing data optimization processing on the multi-dimensional design parameter dataset according to the environmental performance score distribution includes a systematic data processing flow. The environmental performance score distribution is used as the basis for weight allocation, and the distribution is derived from the calculation results of the comprehensive evaluation feature set. The mapping relationship of the weight is defined as a positive correlation function of the score value and the influence coefficient, and the building area with a higher score obtains a larger weight coefficient. This weight design strengthens the contribution of the target building and weakens the interference signal of the surrounding buildings or environment.

[0063] The weight filtering is realized by constructing a three-dimensional weight matrix acting on the original dataset. The matrix dimension is completely matched with the space-time structure of the multi-dimensional design parameter dataset, and each matrix element corresponds to the weight value of a specific spatial position and time node. The filtering operation is simultaneously performed in the frequency domain and the time domain, and the non-uniform data distribution is processed using the space-variant filtering technique. The core parameters of the filter kernel function design include the spatial smoothing radius and the time decay constant, considering the physical continuity of the building structure. The processed dataset significantly reduces the data energy proportion of the non-target area, and retains the main features of the target building.

[0064] The generation of the optimized dataset goes through multiple processing stages. First, the basic weight filtering is performed to suppress the obvious cross-building interference components. Then, adaptive compensation filtering is performed to compensate for the data attenuation in the boundary area of the target building. Finally, band-pass filtering is implemented to retain the typical frequency band data related to building performance evaluation. These three-stage processing ensures effective suppression of interference components while maintaining the integrity and continuity of the target building data. The output dataset uses a hierarchical storage structure, with the spatial dimension partitioned according to building elevation and the time dimension organized in event segments.

[0065] The spatio-temporal feature separation algorithm extracts the core features from the optimized dataset. Based on the signal space projection principle, the algorithm constructs a feature subspace with a specially designed basis vector set representing independent features of building energy consumption and structural response. The algorithm calculation process includes two main steps: orthogonal decomposition and feature projection. In the orthogonal decomposition stage, the coupling effect of spatial layout parameters and material property parameters is separated through geometric constraints. In the feature projection stage, the time series data is projected onto the energy consumption feature axis and the structural feature axis to generate independent state vectors.

[0066] The extraction of the optimized energy consumption feature curve uses the principal component regression method. From the state vectors output by the spatio-temporal feature separation, the three principal component components with the highest correlation with energy consumption parameters are selected. These three components are phase-aligned and energy-normalized to synthesize a continuous time curve. The curve sampling rate remains consistent with the original data, and the sampling density is automatically increased during event triggering periods. The curve data is supplemented with a building usage correction coefficient to eliminate noise fluctuations caused by personnel flow.

[0067] The generation of the optimized structural characteristic curve fuses multi-source structural response data. The basic inputs include displacement sensor data, strain gauge data, and accelerometer data. After time-space feature separation, the structural characteristic components of each data source are first extracted. Then, multi-sensor data fusion is performed, and a confidence weighting method is used to solve the conflict of different data sources. During the fusion process, the results of structural modal identification are constrained to ensure that the output curve meets the building mechanics characteristics. The final generated curve takes the structural displacement value as the main coordinate axis, with stress and acceleration auxiliary coordinate axes.

[0068] The data post-processing link ensures the engineering practicability of the output curve. The curve smoothing process uses an adaptive window algorithm, and the window width is dynamically adjusted according to the curve curvature. The characteristic value normalization process refers to the limit standard in building safety specifications, so that the curve value range is between 0 and 1, with clear engineering implications. Time stamp calibration solves the problem of clock offset of multiple devices, and uses a reference time server for global synchronization. Data integrity verification is achieved through cyclic redundancy check, and the error data segment triggers an automatic repair mechanism.

[0069] The system implementation architecture includes distributed computing nodes and a central processing unit. The weight filtering process is completed in the edge computing node, which uses local cache to reduce data transmission delay. The time-space feature separation algorithm runs in the central processing unit, which calls a multi-core parallel optimization computing library. The final synthesis of the characteristic curve is completed in the visualization module, which supports multi-curve superposition and comparative analysis. The entire processing chain has a breakpoint resume function, and unexpected interruptions can be recovered from the nearest checkpoint.

[0070] The output interface is standardized to the general data format in the engineering field. The optimized energy consumption characteristic curve and the optimized structural characteristic curve are output as vector graphics files with time series labels. The file header information includes building number, data collection date, parameter version, and other metadata. Each curve supports data export and analysis callback functions, and seamlessly integrates with building management systems. The result storage uses differential compression technology to reduce storage space occupation.

[0071] The data caching strategy is designed to address large-scale building group application scenarios. The system deploys distributed cache on the server side, and retains the processing results of the last 72 hours in real time. The client request returns the cache result first, reducing the repeated calculation of resource consumption. The cache update mechanism uses a weighted replacement algorithm, and the data with high frequency access is preferentially retained. The user interface supports curve characteristic value retrieval and historical record tracing functions, which facilitate the longitudinal comparison of building performance evolution trends. The technical route of this implementation clearly distinguishes the data processing and feature extraction stages, and realizes the accurate optimization and extraction of building characteristic parameters through the dynamic weight guidance of environmental performance scores.

[0072] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting; it is not intended to exclude myriad other embodiments of the present application that other inventors can develop based on the same general inventive concepts embodied by the described embodiments. That is, although the present application is described in terms of particular embodiments and illustrative figures, it should be apparent that the scope of the present application is not limited to these specific embodiments.

[0073] While the embodiments of the application have been shown and described herein, it will be understood by those of ordinary skill in the art that various changes, modifications, alternatives, and variations can be made to the embodiments without departing from the spirit and scope of the application, which is defined by the appended claims and their equivalents.

Claims

1. A digital intelligent green building design evaluation method, characterized by: The following steps are involved: Collecting a multi-dimensional design parameter data set of a target building, wherein the multi-dimensional design parameter data set includes spatial layout parameters, material property parameters, energy consumption parameters, and environmental interference event parameters; Separating energy consumption characteristic parameters, structural characteristic parameters, and interference characteristic parameters corresponding to environmental interference events from the multi-dimensional design parameter data set; Establishing dynamic correlation features between the energy consumption characteristic parameters and the structural characteristic parameters, and extracting collaborative adjustment features of the environmental interference event on the energy consumption characteristic parameters and the structural characteristic parameters; Based on the dynamic correlation feature and the collaborative adjustment feature, a comprehensive evaluation feature set is generated to characterize the difference in building sustainability; and an environmental performance score distribution of the building design scheme is calculated based on the comprehensive evaluation feature set; Data optimization processing is performed on the multi-dimensional design parameter data set according to the environmental performance score distribution to separate the optimized energy consumption characteristic curve and the optimized structural characteristic curve of the target building.

2. The digital intelligent green building design evaluation method according to claim 1 is characterized in that: The step of separating energy consumption characteristic parameters, structural characteristic parameters, and interference characteristic parameters corresponding to environmental interference events from the multi-dimensional design parameter data set includes the following steps: A dynamic clustering algorithm is used to perform spatial domain division on the multi-dimensional design parameter data set, construct a main channel of energy consumption characteristics, and retain periodic energy consumption fluctuation components within the range of 0.05-1.2 Hz through frequency band screening; A structural feature auxiliary channel based on adaptive weighted fusion is established, and the Gaussian kernel function is used to perform multi-scale decoupling of the structural response data in the 1.2-5Hz frequency band. The independence index of the energy consumption feature main channel and the structural feature auxiliary channel is evaluated by mutual entropy measurement, the fusion weights between the channels are dynamically adjusted to achieve parameter separation, and the energy consumption feature parameters and the structural feature parameters are output.

3. The digital intelligent green building design evaluation method according to claim 1 is characterized in that: Extracting interference characteristic parameters corresponding to environmental interference events from the multi-dimensional design parameter data set includes the following steps: Performing mutation gradient detection on the time series frames of the multi-dimensional design parameter data set to locate candidate time intervals of environmental interference events and generate event markers; Performing non-stationary feature decomposition on the marker data corresponding to the event marker to extract 1-10 Hz frequency band feature components reflecting external environmental disturbances; Based on the spatiotemporal correlation matrix, local environmental interference and systematic noise are distinguished, and the frequency band characteristic components are integrated to generate an interference intensity time series curve, which is an interference characteristic parameter.

4. The digital intelligent green building design evaluation method according to claim 3 is characterized in that: The step of performing sudden gradient detection on the time series frames of the multi-dimensional design parameter data set, locating candidate time intervals of environmental interference events, and generating event markers includes the following steps: Calculate the gradient energy operator difference sequence of adjacent time series frames through the sliding analysis window; Generate dynamic judgment thresholds based on real-time data fluctuation characteristics; Determining a starting time point and a duration range of an environmental interference event based on the dynamic determination threshold; The candidate time interval is located using the starting time point and the duration range, and an environmental interference event marker is generated.

5. The digital intelligent green building design evaluation method according to claim 3 is characterized in that: The method of distinguishing local environmental interference from systematic noise based on the spatiotemporal correlation matrix and fusing the frequency band characteristic components to generate an interference intensity time series curve includes the following steps: Construct a three-dimensional spatiotemporal correlation map based on the phase changes of multi-source data, and calculate the covariance distance between each spatial node and the reference trajectory; When the cumulative covariance distance of continuous spatial nodes exceeds the dynamic noise threshold, it is determined to be a local environmental interference event and the spatial positioning coordinate set is output; The instantaneous energy distribution of the frequency band characteristic component is aligned with the spatial positioning coordinate set in time and space to generate a time-continuous interference intensity time series curve.

6. The digital intelligent green building design evaluation method according to claim 1 is characterized in that: The step of establishing the dynamic correlation feature between the energy consumption characteristic parameter and the structural characteristic parameter comprises the following steps: Calculate the coefficient of variation of structural response and the energy intensity-structural deformation correlation index within the energy consumption fluctuation period to generate dynamic correlation characteristics; Establish environmental interference-energy consumption coupling criteria, analyze the stability index and structural recovery rate of energy consumption characteristic parameters before and after environmental interference events, and generate coordinated regulation characteristics.

7. The digital intelligent green building design evaluation method according to claim 1 is characterized in that: Before generating the comprehensive evaluation feature set representing the differences in building sustainability, the following steps are also included: Determine whether the duration of the environmental interference event exceeds a preset duration threshold, and detect whether there is overlap of multiple building reflection cross sections; The method of generating a comprehensive evaluation feature set that characterizes differences in building sustainability includes the following steps: If the duration of the environmental interference event does not exceed the preset duration threshold and there is no overlap of multiple building reflection cross-sections, an initial evaluation feature set is generated based on the dynamic association feature and the collaborative adjustment feature, and the initial evaluation feature set is used as the comprehensive evaluation feature set.

8. The digital intelligent green building design evaluation method according to claim 7 is characterized in that: The following steps are also included: If the duration of the environmental interference event exceeds a preset duration threshold, or there is overlap of multiple building reflection cross sections, a dynamic clustering algorithm is used to perform secondary decoupling processing on the energy consumption characteristic parameters and the structural characteristic parameters; Based on the three-dimensional spatiotemporal correlation map, the overlapping areas of the reflection cross sections corresponding to the systematic noise are eliminated, and the processed energy consumption characteristic parameters and the processed structural characteristic parameters are re-extracted; A target dynamic association feature is generated according to the processed energy consumption characteristic parameters, and a target collaborative adjustment feature is generated according to the processed structural characteristic parameters; the target dynamic association feature and the target collaborative adjustment feature are used to update the initial evaluation feature set.

9. The digital intelligent green building design evaluation method according to claim 8 is characterized in that: After the dynamic clustering algorithm is used to perform secondary decoupling processing on the energy consumption characteristic parameters and the structural characteristic parameters, the following steps are also included: Verify the convergence of energy distribution of data after secondary decoupling processing; If the verification is successful, the initial evaluation feature set is modified based on the target dynamic association feature and the target collaborative adjustment feature to generate an updated evaluation feature set, and the updated evaluation feature set is used as the comprehensive evaluation feature set.

10. The digital intelligent green building design evaluation method according to claim 1, characterized in that: The performing of data optimization processing on the multi-dimensional design parameter data set according to the environmental performance score distribution comprises the following steps: performing weight filtering on the multi-dimensional design parameter data set based on the environmental performance score distribution, suppressing energy consumption and structural interference components of non-target buildings, and generating an optimized data set; A spatiotemporal feature separation algorithm is executed on the optimized data set to extract the optimized energy consumption characteristic curve and the optimized structural characteristic curve of the target building.

Images