Low-carbon ecological parameter optimization method and system for curtain wall based on multi-dimensional thermal simulation

Abstract

This invention provides a method and system for optimizing low-carbon ecological parameters of curtain walls based on multi-dimensional thermal simulation. The method specifically includes: acquiring curtain wall structure, orientation, shading, window opening, material carbon emissions, and meteorological boundary parameters; establishing a thermal carbon emission correlation calculation model; calculating heat transfer, solar heat gain, heat storage, natural ventilation, and carbon emissions during construction and operation phases; constructing a thermal ecological state matrix; generating thermal response and life-cycle carbon emission response sequences corresponding to each candidate parameter; obtaining a parameter evaluation matrix; calculating thermal inertia, carbon sensitivity, and phase lag indices based on the evaluation matrix; iteratively optimizing candidate parameters after sorting and grouping; freezing parameter groups that meet both heat flow and carbon emission residual standards, and updating non-compliant groups according to a preset step size; synchronously adjusting the evaluation matrix during iteration until the comprehensive target converges; and outputting the optimal low-carbon ecological parameter optimization result per unit curtain wall area based on the converged multi-dimensional core indices.

Description

Technical Field

[0001] This application belongs to the field of parameter optimization, and in particular relates to a method and system for optimizing low-carbon ecological parameters of curtain walls based on multidimensional thermal simulation. Background Technology

[0002] Curtain wall systems encompass numerous parameters, including construction, orientation, shading, window openings, and materials. These parameters not only determine material carbon emissions during construction but also intertwine with solar radiation, natural ventilation, and envelope heat transfer during operation, collectively impacting a building's total energy consumption and carbon emissions. Matrix operations and optimization algorithms can be used to evaluate the thermal performance and energy efficiency of curtain walls. Existing methods often construct thermal state matrices to simulate heat flow and heat gain under specific meteorological boundaries and utilize conventional iterative algorithms to uniformly search and solve for parameters. Simultaneously, they attempt to combine carbon emission and heat flow analysis, evaluating the energy-saving performance of different parameter combinations based on a comprehensive objective function. However, the optimization process fails to analyze the heterogeneous impact of different parameters on thermal performance and carbon emissions, performing a unified global iteration on all variables. This overlooks the inherent differences in thermal inertia, carbon sensitivity, and phase lag dimensions among the parameters, resulting in high computational redundancy and low convergence efficiency in the optimization process. The iterative strategy lacks a mechanism for controlling parameter groups. During iteration, it cannot assess the correlation between frozen and unfrozen parameter groups based on the convergence state of the parameters and adjust the evaluation matrix in real time. This can easily lead the algorithm to get stuck in local optima when dealing with thermal carbon emission correlation constraints, resulting in deviations in operating energy consumption, thermal stability, and natural ventilation indicators. Consequently, it cannot output the optimal low-carbon ecological parameter results per unit curtain wall area. Summary of the Invention

[0003] According to one aspect of the present invention, a method for optimizing low-carbon ecological parameters of curtain walls based on multidimensional thermal simulation is provided, the method comprising: Acquire curtain wall structural parameters, orientation parameters, shading parameters, window opening parameters, material carbon emission parameters, and meteorological boundary parameters. Based on these parameters, establish a curtain wall thermal carbon emission correlation calculation model to calculate the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emission during operation, and carbon emission during construction. Construct a thermal ecological state matrix. Based on the thermal ecological state matrix, generate the thermal response sequence and life-cycle carbon emission response sequence of the curtain wall system corresponding to each candidate parameter at each time period, and obtain the parameter evaluation matrix. The thermal inertia index, carbon sensitivity index, and phase lag index of each candidate parameter are calculated based on the parameter evaluation matrix. The candidate parameters are sorted and grouped according to the index. Iterative optimization is performed on each parameter group. The heat flux residual, carbon emission residual, and change in comprehensive target value are calculated after each iteration. When the heat flux residual and carbon emission residual of the corresponding parameter group are both less than the preset heat flux convergence threshold and carbon emission convergence threshold, respectively, the parameter group is frozen. The unfrozen parameter groups are updated continuously at a preset step size. After each iteration, the parameter evaluation matrix is ​​adjusted according to the correlation between the frozen parameter group and the unfrozen parameter group until the comprehensive target value converges. Based on the converged operating energy consumption index, indoor thermal stability index, natural ventilation effective duration index and life-cycle carbon emission index, the optimized low-carbon ecological parameters of the curtain wall corresponding to the optimal comprehensive target value per unit curtain wall area are output.

[0004] According to another aspect of the present invention, a low-carbon ecological parameter optimization system for curtain walls based on multidimensional thermal simulation is provided, comprising the following modules: The generation module is used to acquire curtain wall structural parameters, orientation parameters, shading parameters, window opening parameters, material carbon emission parameters, and meteorological boundary parameters. Based on these parameters, a curtain wall thermal carbon emission correlation calculation model is established to calculate the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emission during operation, and carbon emission during construction. A thermal ecological state matrix is ​​constructed, and based on the thermal ecological state matrix, the thermal response sequence and the whole life carbon emission response sequence of the curtain wall system corresponding to each candidate parameter are generated at each time period, and a parameter evaluation matrix is ​​obtained. The update module is used to calculate the thermal inertia index, carbon sensitivity index and phase lag index of each candidate parameter according to the parameter evaluation matrix, sort and group the candidate parameters according to the index, perform iterative optimization for each parameter group, calculate the heat flow residual, carbon emission residual and comprehensive target value change after each iteration, freeze the parameter group when the heat flow residual and carbon emission residual of the corresponding parameter group are both less than the preset heat flow convergence threshold and carbon emission convergence threshold respectively, and continue to update the unfrozen parameter group according to the preset step size; The optimization module is used to adjust the parameter evaluation matrix according to the correlation between the frozen parameter group and the unfrozen parameter group after each iteration until the comprehensive target value converges. Based on the converged operating energy consumption index, indoor thermal stability index, effective natural ventilation duration index, and life-cycle carbon emission index, it outputs the optimization result of the low-carbon ecological parameters of the curtain wall corresponding to the optimal comprehensive target value per unit curtain wall area.

[0005] This invention establishes a thermal carbon emission correlation calculation model by acquiring multi-dimensional curtain wall parameters and meteorological boundary parameters, representing heat transfer, heat gain, natural ventilation volume, and total life cycle carbon emissions, generating a thermal response sequence and a total life cycle carbon emission response sequence for the curtain wall system. Based on the parameter evaluation matrix, thermal inertia, carbon sensitivity, and phase lag indices are extracted to sort, group, and iteratively optimize candidate parameters. A group freezing and preset step-size update mechanism improves the optimization calculation efficiency and solution accuracy in a large parameter space. During iteration, the evaluation matrix is ​​adjusted by considering the correlation between frozen and unfrozen parameter groups to ensure stable convergence of the comprehensive target value. This achieves synergistic optimization of low-carbon ecological parameters for the curtain wall, improving the indoor thermal environment quality of the building while reducing the total life cycle carbon emissions of the curtain wall. Attached Figure Description

[0006] Figure 1 The flowchart shows a method for optimizing low-carbon ecological parameters of curtain walls based on multidimensional thermal simulation. Figure 2 This is a schematic diagram of the 24-hour thermal response time series curve; Figure 3 This is a schematic diagram of the cluster distribution of candidate parameters; Figure 4 This is a diagram showing the performance comparison between different groups. Detailed Implementation

[0007] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0008] Example 1 In Embodiment 1 of the present invention, as Figure 1 As shown, a method for optimizing low-carbon ecological parameters of curtain walls based on multidimensional thermal simulation includes: S1. Obtain the curtain wall construction parameters, orientation parameters, shading parameters, window opening parameters, material carbon emission parameters, and meteorological boundary parameters. Based on these parameters, establish a curtain wall thermal carbon emission correlation calculation model to calculate the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emission during operation, and carbon emission during construction. Construct a thermal ecological state matrix. Based on the thermal ecological state matrix, generate the thermal response sequence and life-cycle carbon emission response sequence of the curtain wall system corresponding to each candidate parameter at each time period, and obtain the parameter evaluation matrix.

[0009] Read the basic dataset exported from the building information model and parse out the number of glass layers, air layer thickness, orientation azimuth angle, shading louver tilt angle, window-to-floor area ratio, carbon emission factors of aluminum profiles and glass, and weather data files for typical meteorological years.

[0010] A thermal carbon emission correlation calculation model for the curtain wall was established using the EnergyPlus calculation engine driven by the eppy library. In the thermal calculations, a differential mesh was defined using the numpy library, and hourly heat transfer of the building envelope was calculated based on Fourier's law of thermal conductivity combined with an explicit finite difference algorithm. The hourly solar heat gain was calculated using the perez sky diffuse model from the pvlib library. The lumped parameter method and the Euler multiplier method were used to calculate the structural heat storage. The natural ventilation volume under the combined effects of thermal pressure and wind pressure was calculated using Bernoulli's equation and empirical formulas for orifice outflow. In the carbon emission calculations, the Life Cycle Assessment (LCA) formula was used. Carbon emissions during the operation phase were set as hourly air conditioning energy consumption multiplied by the grid carbon emission factor, while carbon emissions during the construction phase were set as the sum of the volume of each material multiplied by its density and the corresponding material carbon emission parameters.

[0011] The six calculation results are concatenated along the time and parameter dimensions to construct a thermal ecological state matrix with dimensions equal to the number of time step rows multiplied by the number of feature columns. Indoor temperature and energy consumption values ​​for different curtain wall parameter combinations at each hour are extracted from the thermal ecological state matrix to generate a thermal response sequence. The sum of the corresponding building material carbon emissions and operational carbon emissions is then extracted to generate a life-cycle carbon emission response sequence. The thermal response sequence and the life-cycle carbon emission response sequence are horizontally stacked to obtain a parameter evaluation matrix.

[0012] In some embodiments, the step of establishing a curtain wall thermal carbon emission correlation calculation model based on the parameters, calculating the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emissions during operation, and carbon emissions during construction, and constructing a thermal ecological state matrix includes: Based on the curtain wall construction parameters and meteorological boundary parameters, a structural layer temperature heat transfer analytical model is established to calculate the temperature difference between the inner and outer surfaces of the curtain wall under each control step. Combined with the thermal properties of the materials, the heat transfer and heat storage of the enclosure are calculated. Based on the shading parameters, orientation parameters, and meteorological boundary parameters, the solar heat gain that penetrates the curtain wall and enters the room during each time period is calculated by determining the real-time shading ratio of the shading components and the transmittance of the glass. Based on the window opening parameters and meteorological boundary parameters, the number of air changes under the action of wind pressure and thermal pressure is calculated to obtain the natural ventilation volume. Based on the carbon emission parameters of the materials, the carbon emission base corresponding to the energy consumption of building material production, transportation, and construction throughout the entire life cycle of the curtain wall is extracted, the carbon emission during the construction phase is calculated, and the carbon emission during the operation phase is calculated in combination with the energy consumption attributes of the equipment. The thermal ecological state matrix is ​​constructed by vectorizing the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emissions during operation and construction in the time dimension.

[0013] When establishing the analytical model for temperature heat transfer in the structural layer, a one-dimensional implicit finite difference method was adopted, with the control step size set to 1 hour. The real-time temperature difference of 9℃ between the outdoor (35℃) and indoor (26℃) was extracted, and the heat transfer was calculated based on the aluminum alloy profile and the insulated glass. The heat storage was calculated using the specific heat capacity of glass (840 J / (kg·K)) and its density. The heat storage capacity under temperature fluctuations is obtained as follows: To determine the solar heat gain, the preferred angle of the shading louvers is 0°-90°, such as 45°, and the preferred overhang length is 0.3-0.8m, such as 0.5m, taking into account the solar radiation intensity. With a south-southeast orientation of 15°, the shading ratio was calculated to be 65% using a ray tracing algorithm. Combined with a glass SHGC value of 0.35, the solar heat gain was calculated. Natural ventilation is determined by the effective area of ​​the open windows. With a flow coefficient of 0.65, and considering wind pressure and thermal pressure at an outdoor wind speed of 3 m / s, the calculated air exchange rate is 3.5 times per hour, which translates to a ventilation volume of... .

[0014] In the carbon emission accounting and matrix construction stage, carbon emission factors from glass production are extracted. Aluminum materials and construction factors The total carbon emissions during the construction phase are calculated as follows: During the operation phase, the HVAC COP value is 3.5 and the electricity carbon emission factor is used. The annual carbon emissions were calculated as follows: .

[0015] The extracted heat transfer, heat gain, heat storage, ventilation volume, and two-stage carbon emissions were mapped to the dimensionless interval [0,1] using Min-Max range standardization. Time-series vectors were generated according to time steps t=1 to 8760, and seamlessly spliced ​​column-wise to generate a high-dimensional feature set of N×6, which constitutes the thermal ecological state matrix.

[0016] In some embodiments, generating the thermal response sequence and life-cycle carbon emission response sequence of the curtain wall system corresponding to each candidate parameter at each time period based on the thermal-ecological state matrix, and obtaining the parameter evaluation matrix, includes: The outdoor ambient temperature, indoor surface temperature, indoor air temperature and heat transfer of the curtain wall system corresponding to each candidate parameter are extracted according to the time series to construct a thermal response sequence that changes with time within a 24-hour period. Extract the carbon emissions of each candidate parameter during the material production, construction, and operation stages, as well as the expected service life of the corresponding curtain wall components. Calculate the annual equivalent carbon emissions of each parameter and convert the emissions into equivalent carbon emission rates corresponding to the time step of the thermal response sequence. Expand to generate a full-life carbon emission response sequence. The normalized thermal response sequence and the full-lifetime carbon emission response sequence are used as feature vectors of different dimensions and then concatenated into a matrix to generate a parameter evaluation matrix that includes thermal and carbon emission features.

[0017] Set time resolution Thermal response sequences were constructed by extracting 24-hour diurnal thermal data. The following parameters were recorded: hourly outdoor ambient temperature under typical summer conditions, indoor air temperature set to a constant 26°C, indoor surface temperature fluctuating due to solar radiation, and heat transfer from the building envelope. These four physical quantities were aligned along the time axis to generate a 24×4 diurnal thermal response sequence matrix. The 24-hour thermal response time series changes of the curtain wall system under typical summer conditions are shown below. Figure 2 As shown.

[0018] In one embodiment, the thermal-ecological state matrix may further include outdoor ambient temperature, indoor surface temperature, and indoor air temperature, and the outdoor ambient temperature, indoor surface temperature, and indoor air temperature, as well as the heat transfer of the building envelope, are extracted from the thermal-ecological state matrix. In another embodiment, the outdoor ambient temperature, indoor surface temperature, and indoor air temperature of the curtain wall system corresponding to each candidate parameter are obtained based on the structural layer temperature heat transfer analysis model and meteorological boundary parameters, and the heat transfer of the building envelope is extracted from the thermal-ecological state matrix according to the time series, and a thermal response sequence varying with time over a 24-hour period is jointly constructed.

[0019] When constructing the full lifecycle carbon emission sequence, the expected service life of the curtain wall components is set at 30 years. If the total carbon emissions from the materials and construction phases are... During the operation phase, the average annual carbon emissions are The total carbon emissions over the entire life cycle are Based on this, the annual equivalent carbon emissions were calculated. Converted to an hourly step size, the equivalent carbon emission rate is approximately... Furthermore, by combining the peak and valley distribution characteristics of energy consumption of building daytime equipment, a column vector of length 24 is generated as the whole life cycle carbon emission response sequence.

[0020] The Z-score normalization method is used to eliminate the dimensional differences of Celsius, power density and carbon equivalent. The 24×4 thermal sequence and the 24×1 carbon emission sequence are concatenated by the feature dimension. The 100 sampled candidate curtain wall parameters are combined to generate a three-dimensional parameter evaluation matrix tensor with dimensions of 100×24×5.

[0021] S2, calculate the thermal inertia index, carbon sensitivity index and phase lag index of each candidate parameter according to the parameter evaluation matrix, sort and group the candidate parameters according to the index, perform iterative optimization for each parameter group, calculate the heat flow residual, carbon emission residual and comprehensive target value change after each iteration, freeze the parameter group when the heat flow residual and carbon emission residual of the corresponding parameter group are both less than the preset heat flow convergence threshold and carbon emission convergence threshold respectively, and continue to update the unfrozen parameter group according to the preset step size.

[0022] The `gradient` function is used to calculate the time derivative of the thermal response sequence in the parameter evaluation matrix, thus determining the thermal inertia index representing the degree of temperature decay. The `pearsonr` function is used to calculate the Pearson correlation coefficient between the candidate parameter's small perturbation and the lifetime carbon emission response sequence to determine the carbon sensitivity index. A Fast Fourier Transform (FFT) is performed on the thermal response sequence and the outdoor meteorological temperature sequence, using the `fft` module of the NumPy library to extract the phase difference as the phase lag index. Based on the calculated three indices, a clustering model is established using the KMeans clustering algorithm, and the number of clusters is set to divide the candidate parameters into a high-frequency sensitive group, a mid-frequency slowly varying group, and a low-frequency solid-state group.

[0023] A block coordinate descent optimization algorithm was written in Python to perform iterative optimization on each parameter set. After each iteration, the root mean square error (RMSE) calculation function from the sklearn library was used to calculate the root mean square error between the predicted heat flux value of the current iteration and the predicted life-cycle carbon emissions value of the previous iteration, as the heat flux residual. The root mean square error between the predicted life-cycle carbon emissions value of the current iteration and the predicted life-cycle carbon emissions value of the previous iteration was also calculated as the carbon emissions residual. The overall target value was calculated by weighted summing of total building energy consumption and total carbon emissions in the objective function, and the change in the overall target value was obtained by calculating the absolute difference of this weighted sum between two adjacent iterations. For example, if the energy consumption weight is set to 0.7 and the carbon emission weight to 0.3, and a certain parameter combination predicts a building's total energy consumption of 100 units and total carbon emissions of 50 units, the comprehensive target value for this iteration is calculated to be 85. If, for a specific building with a strong emphasis on carbon reduction requirements, the energy consumption weight is adjusted to 0.4 and the carbon emission weight is set to 0.6, and the predicted total energy consumption is 120 units and total carbon emissions are 40 units, then the comprehensive target value for this round is calculated to be 72. When the units for total building energy consumption and total carbon emissions are inconsistent, this can be achieved through normalization or by setting different units for the weights. This is obvious to those skilled in the art and will not be elaborated further.

[0024] Set a conditional statement to check if the residual heat flux of the high-frequency sensitive group or the medium-frequency slowly varying group is less than a preset value. And the carbon emission residual is less than the preset value. When the condition is met, the state of the parameter group is marked as frozen using a Boolean index. For unfrozen parameter groups that do not meet the above convergence threshold, the Adam algorithm, a moment estimation optimizer, is used to calculate the gradient descent direction and update amount with a decaying learning rate as the preset step size, and the original parameter values ​​are added for the next round of updates.

[0025] In some embodiments, the step of calculating the thermal inertia index, carbon sensitivity index, and phase hysteresis index of each candidate parameter according to the parameter evaluation matrix, and sorting and grouping the candidate parameters according to the index, includes: The peak times of outdoor ambient temperature and indoor side surface temperature are extracted from the thermal response sequence, and the absolute time difference between the two is calculated as the phase lag index. For each candidate parameter, a perturbation variable with a preset step size is set, and the relative change rate of the whole life carbon emission response sequence before and after the perturbation is calculated as the carbon sensitivity index of the candidate parameter. The thermal inertia index is calculated by combining the thickness, thermal conductivity, density, and specific heat capacity of the corresponding material; Cluster analysis is performed on each candidate parameter based on the thermal inertia index, carbon sensitivity index, and phase hysteresis index. Candidate parameters with similar characteristics are grouped into the same parameter group, thus completing the sorting and grouping of candidate parameters.

[0026] When extracting the phase lag index, a peak detection algorithm is used to traverse the 24-hour thermal response sequence to locate the peak time of the outdoor ambient temperature in summer, such as 14:30 (absolute time), and the peak time of the indoor surface temperature under the dual delay of curtain wall thermal resistance and heat capacity, such as 17:00. The absolute time difference is calculated as 2.5 hours using the difference in timestamps, and this is used as the phase lag index for this parameter. For continuous parameters of the glass heat transfer coefficient, a uniform perturbation step of ±5% of the nominal value is set. It is assumed that the glass U-value is perturbed from 1.8 to... At that time, it triggered a surge in total life-cycle carbon emissions from 450 to The relative increase was 2.67%, and the calculated quantification value of the carbon sensitivity index for this parameter was 0.534.

[0027] To calculate the thermal inertia index D, the standard formula D=R×S is used. Taking the parameters of the built-in rock wool insulation layer as an example, parameters of thickness 0.05m, thermal conductivity 0.04W / (m·K), and density are used. Given a specific heat capacity of 800 J / (kg·K), the product of thermal resistance R and heat storage coefficient S was calculated, yielding a thermal inertia index value of 1.2. After extraction, the three core parameters were encapsulated into a three-dimensional feature vector [2.5, 0.534, 1.2], and mapped to a high-dimensional Euclidean feature space using the K-Means++ clustering algorithm. With a fixed cluster size K=4, multiple iterations were performed to divide the parameter sets into groups such as high-sensitivity low-inertia and low-sensitivity high-inertia groups. Within each subset, a descending order was forcibly performed according to the thermal inertia index, completing the dimensionality reduction and cost reduction grouping preparation for the optimization task. The clustering distribution results of each candidate parameter based on the carbon sensitivity index and phase lag index are as follows: Figure 3 As shown.

[0028] In some embodiments, the iterative optimization of each parameter group, calculating the heat flux residual, carbon emission residual, and change in the comprehensive target value after each iteration, and freezing the parameter group when the heat flux residual and carbon emission residual of the corresponding parameter group are simultaneously less than the preset heat flux convergence threshold and carbon emission convergence threshold, respectively, and continuing to update the unfrozen parameter groups according to a preset step size, includes: In two adjacent iterations, the total heat transfer loss and total carbon emissions of the curtain wall system affected by the current parameter group variable update are extracted respectively. The absolute difference between the corresponding values ​​of the current round and the previous round is calculated and recorded as heat flow residual and carbon emission residual respectively. The heat flux residual and the carbon emission residual are compared with preset heat flux convergence thresholds and carbon emission convergence thresholds, respectively. If the heat flux residual and carbon emission residual of a parameter group are both less than the corresponding convergence threshold, the current parameter value of the parameter group is written to the static repository and frozen, and will not be updated in subsequent iterations. If the values ​​are not simultaneously less than the corresponding convergence thresholds, the variables in the parameter group are updated according to the preset optimization gradient step size, and the next iteration begins.

[0029] In two consecutive rounds of parallel optimization simulations, the overall building status data affected by parameter updates is captured in real time. Assume the total global heat transfer loss calculated in the nth iteration is 1250 kWh / year, and the total carbon emissions are... After the (n+1)th optimization, the energy consumption was reduced to 1248.5 kWh / year and... The heat flux residual obtained through first-order subtraction is 1.5 kWh / year, and the carbon emission residual is... At this point, a pre-configured heat flux convergence threshold, typically set to 0.1% of the initial total heat flux (e.g., 1.2 kWh / year), and a carbon emission convergence threshold, such as the absolute boundary value, are extracted. The calculated residual vector [1.5, 0.8] is then compared with the threshold matrix [1.2, 0.5] using a Boolean cross-match.

[0030] Since the measured residuals did not simultaneously break through the lower convergence limit, the miss mechanism was triggered, rejecting the parameter freezing instruction and activating the optimization gradient module for the unfrozen parameter set. Configure the initial learning rate. As the gradient step size for optimization, the gradient decay direction is obtained by calculating the partial derivative of the objective function with respect to the shading length variable, and then executed. The step size is updated, for example, extending the cantilever length from 0.5m to 0.52m. When the algorithm extrapolates to the (n+k)th round, the heat flow residual is reduced to 0.9kWh / year and the carbon emission residual is reduced to... The condition matrix dual-channel verification both passed. At this point, an SQL command was invoked to write [0.52m] and the bound U value to the static storage pool of the Redis cache to freeze the data and remove redundant calculations.

[0031] S3. After each iteration, the parameter evaluation matrix is ​​adjusted according to the correlation between the frozen parameter group and the unfrozen parameter group until the comprehensive target value converges. Based on the converged operating energy consumption index, indoor thermal stability index, natural ventilation effective duration index and full life carbon emission index, the optimized result of low-carbon ecological parameters of the curtain wall corresponding to the optimal comprehensive target value per unit curtain wall area is output.

[0032] After each update of the unfrozen parameter set by the Adam optimizer, the partial correlation coefficients between the frozen and unfrozen parameter set variables are calculated using the statsmodels library, constructing a cross-influence Jacobian matrix. This Jacobian matrix is ​​multiplied by the update step size of the unfrozen parameters to obtain a state compensation vector. The `add` function from the NumPy library is then used to add this state compensation vector to the corresponding positions in the original parameter evaluation matrix, completing matrix adjustment and correcting the bias caused by parameter correlation.

[0033] The termination condition for the outer loop is set to the change in the overall objective value for five consecutive iterations being less than [a certain value]. When this condition is met, the algorithm is deemed to have converged to the comprehensive target value. After convergence, the algorithm extracts the annual cumulative air conditioning power consumption from the updated parameter evaluation matrix as the operating energy consumption index, calculates the percentage of time the indoor temperature is within the human thermal comfort range as the indoor thermal stability index, counts the total number of hours the windows are opened to meet the ventilation standards as the natural ventilation duration index, and extracts the total carbon emissions from construction and operation as the life-cycle carbon emission index.

[0034] The sTOPSIS multi-criteria decision analysis algorithm is invoked, and the four indicators are assigned weights of 30%, 20%, 20%, and 30%, respectively. The relative closeness of each candidate parameter combination to the ideal optimal solution is calculated. The argmax function of the NumPy library is used to index the row with the largest relative closeness value. The specific numerical combination of curtain wall construction parameters, orientation parameters, shading parameters, and window opening parameters corresponding to that row is exported through the built-in print function and local file writing operations. This is the optimization result of the low-carbon ecological parameters of the curtain wall corresponding to the optimal comprehensive target value per unit curtain wall area.

[0035] In some embodiments, adjusting the parameter evaluation matrix after each iteration based on the correlation between the frozen parameter group and the unfrozen parameter group until the comprehensive target value converges includes: Extract the current state values ​​of each fixed variable in the frozen parameter group, and calculate the correlation weights of the current state values ​​with the correlation weights of the thermal conductivity coefficients and carbon emissions of each active variable in the unfrozen parameter group. The correlation between the thermal conductivity coefficient and the carbon emission correlation weight is used as a feedback correction term to perform a weighted correction on the original corresponding elements in the parameter evaluation matrix, thereby completing the adjustment of the parameter evaluation matrix. Based on the adjusted parameter evaluation matrix, the iterative optimization is re-executed, and the comprehensive target value generated after each iteration is monitored. Calculate the change between the current round's comprehensive target value and the previous round's comprehensive target value. When the change is consistently less than a preset convergence threshold for multiple rounds, the comprehensive target value is determined to have reached a convergence state.

[0036] When a frozen event is detected where the outer shading angle is fixed at 45° and the extension length is locked at 0.5m, the solidification constant parameters are extracted, and a multivariate partial correlation analysis function is used to construct an evaluation surface for the impact on the parameters of the unfrozen glass. Calculations using the correlation model indicate that this fixed shading configuration effectively shields and cancels out the heat gain of the internal glass, with an output correlation thermal conductivity coefficient of 0.85. Simultaneously, based on the implicit energy consumption relationship, the carbon emission correlation weight of the material itself is calculated and assigned a value of 0.92. The range [0.85, 0.92] is established as the rigid feedback correction multiplier for adjusting the bottom layer of the matrix.

[0037] Entering the data flow transformation phase, the corresponding row vectors of the active variables that have not yet been frozen are indexed within the 24×5 parameter evaluation matrix in memory. The original thermal response dimension column data is forcibly multiplied by 1×0.85, and all lifecycle carbon emission characteristic dimension column data are multiplied by 0.92. After completing the underlying parameter tuning through this asymmetric penalty weighting mechanism, a new round of evolutionary optimization calculation is restarted based on the repaired evaluation matrix. Using four core indicators—operating energy consumption, thermal stability, natural ventilation duration, and lifecycle carbon emissions—according to a fixed ratio coefficient of [0.4, 0.2, 0.1, 0.3], a comprehensive target value for each individual component is calculated, such as 0.765. A monitoring queue is deployed to extract the absolute changes in continuous evolution. The sequence has a strict global convergence criterion. During rounds 50 to 55, the monitoring instrument captured five sets of evolution increments [0.00008, 0.00005, 0.00009, 0.00003, 0.00001]. Since the decay sequence of these five consecutive rounds was strictly constrained below the limit threshold, the overall control platform officially issued a stop beacon, determined that the multi-objective solver had reached the optimal convergence state, and output the optimal low-carbon curtain wall construction scheme.

[0038] Using a standard office building model as the test object, simulation calculations were performed using typical summer meteorological data and initial curtain wall structural parameters. The experiment was divided into two batches: a conventional control group and a core experimental group. The conventional control group employed a global multi-objective optimization algorithm, where all curtain wall variables evolved synchronously throughout the entire calculation cycle without triggering group freezing or bottom-level weight correction. The core experimental group deployed the proposed iterative optimization architecture, utilizing a parameter solidification strategy based on heat flux residuals and carbon emission convergence thresholds. After each iteration, the associated thermal conductivity coefficients of the frozen variables were extracted to perform feedback correction on the parameter evaluation matrix.

[0039] The conventional control group only gradually stabilized its overall target value after 320 rounds of computation, with a total heat transfer loss of up to 1315 kWh / year and total carbon emissions over its entire lifespan remaining at [missing information]. In contrast, the core experimental group's evolution increments fell below the preset global convergence threshold after only five consecutive iterations by the 55th round. The optimal construction scheme established by this group controlled the total heat transfer loss to 1248.5 kWh / year, while simultaneously reducing the total carbon emissions over its entire lifecycle to a minimum. A comprehensive comparison of iterative efficiency, thermal performance, and carbon emission performance between the conventional control group and the core experimental group, such as... Figure 4 As shown.

[0040] The core experimental group reduced the number of computational iterations by over 82%, optimized total heat transfer loss by 5.0%, and reduced lifetime carbon emissions by 2.9%. This performance leap is attributed to the data freezing strategy of the static storage pool, which prevents repeated and ineffective derivative calculations of redundant variables, thereby reducing the optimization cost in the high-dimensional feature space. Furthermore, the asymmetric weighted intervention mechanism of the correlation matrix maps the shielding effect between the external shading entity and the inner light-transmitting material, avoiding the index cancellation trap often occurring in conventional global optimization, and achieving simultaneous compliance with the thermal and low-carbon characteristics of the building envelope.

[0041] Example 2 Embodiment 2 of the present invention proposes a low-carbon ecological parameter optimization system for curtain walls based on multi-dimensional thermal simulation, comprising the following modules: The generation module is used to acquire curtain wall structural parameters, orientation parameters, shading parameters, window opening parameters, material carbon emission parameters, and meteorological boundary parameters. Based on these parameters, a curtain wall thermal carbon emission correlation calculation model is established to calculate the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emission during operation, and carbon emission during construction. A thermal ecological state matrix is ​​constructed, and based on the thermal ecological state matrix, the thermal response sequence and the whole life carbon emission response sequence of the curtain wall system corresponding to each candidate parameter are generated at each time period, and a parameter evaluation matrix is ​​obtained. The update module is used to calculate the thermal inertia index, carbon sensitivity index and phase lag index of each candidate parameter according to the parameter evaluation matrix, sort and group the candidate parameters according to the index, perform iterative optimization for each parameter group, calculate the heat flow residual, carbon emission residual and comprehensive target value change after each iteration, freeze the parameter group when the heat flow residual and carbon emission residual of the corresponding parameter group are both less than the preset heat flow convergence threshold and carbon emission convergence threshold respectively, and continue to update the unfrozen parameter group according to the preset step size; The optimization module is used to adjust the parameter evaluation matrix according to the correlation between the frozen parameter group and the unfrozen parameter group after each iteration until the comprehensive target value converges. Based on the converged operating energy consumption index, indoor thermal stability index, effective natural ventilation duration index, and life-cycle carbon emission index, it outputs the optimization result of the low-carbon ecological parameters of the curtain wall corresponding to the optimal comprehensive target value per unit curtain wall area.

[0042] In some embodiments, the step of establishing a curtain wall thermal carbon emission correlation calculation model based on the parameters, calculating the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emissions during operation, and carbon emissions during construction, and constructing a thermal ecological state matrix includes: Based on the curtain wall construction parameters and meteorological boundary parameters, a structural layer temperature heat transfer analytical model is established to calculate the temperature difference between the inner and outer surfaces of the curtain wall under each control step. Combined with the thermal properties of the materials, the heat transfer and heat storage of the enclosure are calculated. Based on the shading parameters, orientation parameters, and meteorological boundary parameters, the solar heat gain that penetrates the curtain wall and enters the room during each time period is calculated by determining the real-time shading ratio of the shading components and the transmittance of the glass. Based on the window opening parameters and meteorological boundary parameters, the number of air changes under the action of wind pressure and thermal pressure is calculated to obtain the natural ventilation volume. Based on the carbon emission parameters of the materials, the carbon emission base corresponding to the energy consumption of building material production, transportation, and construction throughout the entire life cycle of the curtain wall is extracted, the carbon emission during the construction phase is calculated, and the carbon emission during the operation phase is calculated in combination with the energy consumption attributes of the equipment. The thermal ecological state matrix is ​​constructed by vectorizing the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emissions during operation and construction in the time dimension.

[0043] In some embodiments, generating the thermal response sequence and life-cycle carbon emission response sequence of the curtain wall system corresponding to each candidate parameter at each time period based on the thermal-ecological state matrix, and obtaining the parameter evaluation matrix, includes: The outdoor ambient temperature, indoor surface temperature, indoor air temperature and heat transfer of the curtain wall system corresponding to each candidate parameter are extracted according to the time series to construct a thermal response sequence that changes with time within a 24-hour period. Extract the carbon emissions of each candidate parameter during the material production, construction, and operation stages, as well as the expected service life of the corresponding curtain wall components. Calculate the annual equivalent carbon emissions of each parameter and convert the emissions into equivalent carbon emission rates corresponding to the time step of the thermal response sequence. Expand to generate a full-life carbon emission response sequence. The normalized thermal response sequence and the full-lifetime carbon emission response sequence are used as feature vectors of different dimensions and then concatenated into a matrix to generate a parameter evaluation matrix that includes thermal and carbon emission features.

[0044] In some embodiments, the step of calculating the thermal inertia index, carbon sensitivity index, and phase hysteresis index of each candidate parameter according to the parameter evaluation matrix, and sorting and grouping the candidate parameters according to the index, includes: The peak times of outdoor ambient temperature and indoor side surface temperature are extracted from the thermal response sequence, and the absolute time difference between the two is calculated as the phase lag index. For each candidate parameter, a perturbation variable with a preset step size is set, and the relative change rate of the whole life carbon emission response sequence before and after the perturbation is calculated as the carbon sensitivity index of the candidate parameter. The thermal inertia index is calculated by combining the thickness, thermal conductivity, density, and specific heat capacity of the corresponding material; Cluster analysis is performed on each candidate parameter based on the thermal inertia index, carbon sensitivity index, and phase hysteresis index. Candidate parameters with similar characteristics are grouped into the same parameter group, thus completing the sorting and grouping of candidate parameters.

[0045] In some embodiments, the iterative optimization of each parameter group, calculating the heat flux residual, carbon emission residual, and change in the comprehensive target value after each iteration, and freezing the parameter group when the heat flux residual and carbon emission residual of the corresponding parameter group are simultaneously less than the preset heat flux convergence threshold and carbon emission convergence threshold, respectively, and continuing to update the unfrozen parameter groups according to a preset step size, includes: In two adjacent iterations, the total heat transfer loss and total carbon emissions of the curtain wall system affected by the current parameter group variable update are extracted respectively. The absolute difference between the corresponding values ​​of the current round and the previous round is calculated and recorded as heat flow residual and carbon emission residual respectively. The heat flux residual and the carbon emission residual are compared with preset heat flux convergence thresholds and carbon emission convergence thresholds, respectively. If the heat flux residual and carbon emission residual of a parameter group are both less than the corresponding convergence threshold, the current parameter value of the parameter group is written to the static repository and frozen, and will not be updated in subsequent iterations. If the values ​​are not simultaneously less than the corresponding convergence thresholds, the variables in the parameter group are updated according to the preset optimization gradient step size, and the next iteration begins.

[0046] In some embodiments, adjusting the parameter evaluation matrix after each iteration based on the correlation between the frozen parameter group and the unfrozen parameter group until the comprehensive target value converges includes: Extract the current state values ​​of each fixed variable in the frozen parameter group, and calculate the correlation weights of the current state values ​​with the correlation weights of the thermal conductivity coefficients and carbon emissions of each active variable in the unfrozen parameter group. The correlation between the thermal conductivity coefficient and the carbon emission correlation weight is used as a feedback correction term to perform a weighted correction on the original corresponding elements in the parameter evaluation matrix, thereby completing the adjustment of the parameter evaluation matrix. Based on the adjusted parameter evaluation matrix, the iterative optimization is re-executed, and the comprehensive target value generated after each iteration is monitored. Calculate the change between the current round's comprehensive target value and the previous round's comprehensive target value. When the change is consistently less than a preset convergence threshold for multiple rounds, the comprehensive target value is determined to have reached a convergence state.

[0047] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0048] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for optimizing low-carbon ecological parameters of curtain walls based on multidimensional thermal simulation, characterized in that, include: Acquire curtain wall structural parameters, orientation parameters, shading parameters, window opening parameters, material carbon emission parameters, and meteorological boundary parameters. Based on these parameters, establish a curtain wall thermal carbon emission correlation calculation model to calculate the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emission during operation, and carbon emission during construction. Construct a thermal ecological state matrix. Based on the thermal ecological state matrix, generate the thermal response sequence and life-cycle carbon emission response sequence of the curtain wall system corresponding to each candidate parameter at each time period, and obtain the parameter evaluation matrix. The thermal inertia index, carbon sensitivity index, and phase lag index of each candidate parameter are calculated based on the parameter evaluation matrix. The candidate parameters are sorted and grouped according to the index. Iterative optimization is performed on each parameter group. The heat flux residual, carbon emission residual, and change in comprehensive target value are calculated after each iteration. When the heat flux residual and carbon emission residual of the corresponding parameter group are both less than the preset heat flux convergence threshold and carbon emission convergence threshold, respectively, the parameter group is frozen. The unfrozen parameter groups are updated continuously at a preset step size. After each iteration, the parameter evaluation matrix is ​​adjusted according to the correlation between the frozen parameter group and the unfrozen parameter group until the comprehensive target value converges. Based on the converged operating energy consumption index, indoor thermal stability index, natural ventilation effective duration index and life-cycle carbon emission index, the optimized low-carbon ecological parameters of the curtain wall corresponding to the optimal comprehensive target value per unit curtain wall area are output.

2. The method according to claim 1, characterized in that, The aforementioned parameter-based model for calculating the thermal carbon emissions of the curtain wall is used to calculate the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emissions during operation, and carbon emissions during construction, thereby constructing a thermal ecological state matrix, including: Based on the curtain wall construction parameters and meteorological boundary parameters, a structural layer temperature heat transfer analytical model is established to calculate the temperature difference between the inner and outer surfaces of the curtain wall under each control step. Combined with the thermal properties of the materials, the heat transfer and heat storage of the enclosure are calculated. Based on the shading parameters, orientation parameters, and meteorological boundary parameters, the solar heat gain that penetrates the curtain wall and enters the room during each time period is calculated by determining the real-time shading ratio of the shading components and the transmittance of the glass. Based on the window opening parameters and meteorological boundary parameters, the number of air changes under the action of wind pressure and thermal pressure is calculated to obtain the natural ventilation volume. Based on the carbon emission parameters of the materials, the carbon emission base corresponding to the energy consumption of building material production, transportation, and construction throughout the entire life cycle of the curtain wall is extracted, the carbon emission during the construction phase is calculated, and the carbon emission during the operation phase is calculated in combination with the energy consumption attributes of the equipment. The thermal ecological state matrix is ​​constructed by vectorizing the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emissions during operation and construction in the time dimension.

3. The method according to claim 1, characterized in that, The process involves generating the thermal response sequence and life-cycle carbon emission response sequence of the curtain wall system corresponding to each candidate parameter at each time period based on the thermal-ecological state matrix, and obtaining the parameter evaluation matrix, including: The outdoor ambient temperature, indoor surface temperature, indoor air temperature and heat transfer of the curtain wall system corresponding to each candidate parameter are extracted according to the time series to construct a thermal response sequence that changes with time within a 24-hour period. Extract the carbon emissions of each candidate parameter during the material production, construction, and operation stages, as well as the expected service life of the corresponding curtain wall components. Calculate the annual equivalent carbon emissions of each parameter and convert the emissions into equivalent carbon emission rates corresponding to the time step of the thermal response sequence. Expand to generate a full-life carbon emission response sequence. The normalized thermal response sequence and the full-lifetime carbon emission response sequence are used as feature vectors of different dimensions and then concatenated into a matrix to generate a parameter evaluation matrix that includes thermal and carbon emission features.

4. The method according to claim 1, characterized in that, The step of calculating the thermal inertia index, carbon sensitivity index, and phase hysteresis index of each candidate parameter based on the parameter evaluation matrix, and sorting and grouping the candidate parameters according to the index, includes: The peak times of outdoor ambient temperature and indoor side surface temperature are extracted from the thermal response sequence, and the absolute time difference between the two is calculated as the phase lag index. For each candidate parameter, a perturbation variable with a preset step size is set, and the relative change rate of the whole life carbon emission response sequence before and after the perturbation is calculated as the carbon sensitivity index of the candidate parameter. The thermal inertia index is calculated by combining the thickness, thermal conductivity, density, and specific heat capacity of the corresponding material; Cluster analysis is performed on each candidate parameter based on the thermal inertia index, carbon sensitivity index, and phase hysteresis index. Candidate parameters with similar characteristics are grouped into the same parameter group, thus completing the sorting and grouping of candidate parameters.

5. The method according to claim 1, characterized in that, The iterative optimization of each parameter group involves calculating the heat flux residual, carbon emission residual, and change in the comprehensive target value after each iteration. When the heat flux residual and carbon emission residual of the corresponding parameter group are simultaneously less than the preset heat flux convergence threshold and carbon emission convergence threshold, respectively, the parameter group is frozen. Unfrozen parameter groups are then continuously updated according to a preset step size, including: In two adjacent iterations, the total heat transfer loss and total carbon emissions of the curtain wall system affected by the current parameter group variable update are extracted respectively. The absolute difference between the corresponding values ​​of the current round and the previous round is calculated and recorded as heat flow residual and carbon emission residual respectively. The heat flux residual and the carbon emission residual are compared with preset heat flux convergence thresholds and carbon emission convergence thresholds, respectively. If the heat flux residual and carbon emission residual of a parameter group are both less than the corresponding convergence threshold, the current parameter value of the parameter group is written to the static repository and frozen, and will not be updated in subsequent iterations. If the values ​​are not simultaneously less than the corresponding convergence thresholds, the variables in the parameter group are updated according to the preset optimization gradient step size, and the next iteration begins.

6. The method according to claim 1, characterized in that, The step of adjusting the parameter evaluation matrix after each iteration based on the correlation between the frozen parameter group and the unfrozen parameter group until the comprehensive target value converges includes: Extract the current state values ​​of each fixed variable in the frozen parameter group, and calculate the correlation weights of the current state values ​​with the correlation weights of the thermal conductivity coefficients and carbon emissions of each active variable in the unfrozen parameter group. The correlation between the thermal conductivity coefficient and the carbon emission correlation weight is used as a feedback correction term to perform a weighted correction on the original corresponding elements in the parameter evaluation matrix, thereby completing the adjustment of the parameter evaluation matrix. Based on the adjusted parameter evaluation matrix, the iterative optimization is re-executed, and the comprehensive target value generated after each iteration is monitored. Calculate the change between the current round's comprehensive target value and the previous round's comprehensive target value. When the change is consistently less than a preset convergence threshold for multiple rounds, the comprehensive target value is determined to have reached a convergence state.

7. A curtain wall low-carbon ecological parameter optimization system based on multi-dimensional thermal simulation, characterized in that, Includes the following modules: The generation module is used to acquire curtain wall structural parameters, orientation parameters, shading parameters, window opening parameters, material carbon emission parameters, and meteorological boundary parameters. Based on these parameters, a curtain wall thermal carbon emission correlation calculation model is established to calculate the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emission during operation, and carbon emission during construction. A thermal ecological state matrix is ​​constructed, and based on the thermal ecological state matrix, the thermal response sequence and the whole life carbon emission response sequence of the curtain wall system corresponding to each candidate parameter are generated at each time period, and a parameter evaluation matrix is ​​obtained. The update module is used to calculate the thermal inertia index, carbon sensitivity index and phase lag index of each candidate parameter according to the parameter evaluation matrix, sort and group the candidate parameters according to the index, perform iterative optimization for each parameter group, calculate the heat flow residual, carbon emission residual and comprehensive target value change after each iteration, freeze the parameter group when the heat flow residual and carbon emission residual of the corresponding parameter group are both less than the preset heat flow convergence threshold and carbon emission convergence threshold respectively, and continue to update the unfrozen parameter group according to the preset step size; The optimization module is used to adjust the parameter evaluation matrix according to the correlation between the frozen parameter group and the unfrozen parameter group after each iteration until the comprehensive target value converges. Based on the converged operating energy consumption index, indoor thermal stability index, effective natural ventilation duration index, and life-cycle carbon emission index, it outputs the optimization result of the low-carbon ecological parameters of the curtain wall corresponding to the optimal comprehensive target value per unit curtain wall area.

8. The curtain wall low-carbon ecological parameter optimization system according to claim 7, characterized in that, The aforementioned parameter-based model for calculating the thermal carbon emissions of the curtain wall is used to calculate the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emissions during operation, and carbon emissions during construction, thereby constructing a thermal ecological state matrix, including: Based on the curtain wall construction parameters and meteorological boundary parameters, a structural layer temperature heat transfer analytical model is established to calculate the temperature difference between the inner and outer surfaces of the curtain wall under each control step. Combined with the thermal properties of the materials, the heat transfer and heat storage of the enclosure are calculated. Based on the shading parameters, orientation parameters, and meteorological boundary parameters, the solar heat gain that penetrates the curtain wall and enters the room during each time period is calculated by determining the real-time shading ratio of the shading components and the transmittance of the glass. Based on the window opening parameters and meteorological boundary parameters, the number of air changes under the action of wind pressure and thermal pressure is calculated to obtain the natural ventilation volume. Based on the carbon emission parameters of the materials, the carbon emission base corresponding to the energy consumption of building material production, transportation, and construction throughout the entire life cycle of the curtain wall is extracted, the carbon emission during the construction phase is calculated, and the carbon emission during the operation phase is calculated in combination with the energy consumption attributes of the equipment. The thermal ecological state matrix is ​​constructed by vectorizing the heat transfer of the building envelope, solar heat gain, heat storage, natural ventilation, carbon emissions during operation and construction in the time dimension.

9. The curtain wall low-carbon ecological parameter optimization system according to claim 7, characterized in that, The process involves generating the thermal response sequence and life-cycle carbon emission response sequence of the curtain wall system corresponding to each candidate parameter at each time period based on the thermal-ecological state matrix, and obtaining the parameter evaluation matrix, including: The outdoor ambient temperature, indoor surface temperature, indoor air temperature and heat transfer of the curtain wall system corresponding to each candidate parameter are extracted according to the time series to construct a thermal response sequence that changes with time within a 24-hour period. Extract the carbon emissions of each candidate parameter during the material production, construction, and operation stages, as well as the expected service life of the corresponding curtain wall components. Calculate the annual equivalent carbon emissions of each parameter and convert the emissions into equivalent carbon emission rates corresponding to the time step of the thermal response sequence. Expand to generate a full-life carbon emission response sequence. The normalized thermal response sequence and the full-lifetime carbon emission response sequence are used as feature vectors of different dimensions and then concatenated into a matrix to generate a parameter evaluation matrix that includes thermal and carbon emission features.

10. The curtain wall low-carbon ecological parameter optimization system according to claim 7, characterized in that, The step of calculating the thermal inertia index, carbon sensitivity index, and phase hysteresis index of each candidate parameter based on the parameter evaluation matrix, and sorting and grouping the candidate parameters according to the index, includes: The peak times of outdoor ambient temperature and indoor side surface temperature are extracted from the thermal response sequence, and the absolute time difference between the two is calculated as the phase lag index. For each candidate parameter, a perturbation variable with a preset step size is set, and the relative change rate of the whole life carbon emission response sequence before and after the perturbation is calculated as the carbon sensitivity index of the candidate parameter. The thermal inertia index is calculated by combining the thickness, thermal conductivity, density, and specific heat capacity of the corresponding material; Cluster analysis is performed on each candidate parameter based on the thermal inertia index, carbon sensitivity index, and phase hysteresis index. Candidate parameters with similar characteristics are grouped into the same parameter group, thus completing the sorting and grouping of candidate parameters.

Images