CN112818458B - A building green performance design optimization method and system - Google Patents

Abstract

The invention discloses a method and system for designing and optimizing the green performance of a building, which includes selecting evaluation indicators for evaluating the green performance of the building and determining independent variable parameters that affect the green performance of the building; constructing a building model to be optimized, and importing the building model to be optimized In modeFRONTIER software; according to the selected evaluation indicators for evaluating the green performance of buildings, use modeFRONTIER software to execute a multi-objective optimization algorithm to obtain the Pareto optimization solution set; filter the Pareto optimization solution set to obtain the building green performance design optimization results; this The invention uses modular and visual programming languages, has a simple operation process, and effectively improves the optimization efficiency and the accuracy of the optimal solution set; by performing cluster analysis or multi-criteria decision analysis on the Pareto optimal solution set, the Pareto optimal solution set is realized Screening improves the decision-making efficiency and accuracy of multi-objective optimization design of buildings, assists designers in making design decisions, and improves the objectivity and scientificity of decision-making.

Description

A building green performance design optimization method and system

Technical field

The invention belongs to the technical field of architectural design optimization, and particularly relates to a building green performance design optimization method and system.

Background technique

In the architectural design process, using the optimization of the design plan to balance the green performance of the building is the original intention of the designer to design green buildings; with the development of science and technology, the method of optimizing the green performance of the building using performance simulation and algorithm optimization is applied to architectural design; However, since the Pareto optimal solution of multi-objective optimization is high-dimensional data with multiple design parameters (independent variables) and multiple green performances (dependent variables), and it is difficult to artificially compare the pros and cons of a set of Pareto optimal solutions, Therefore, design decisions are difficult to make. Existing research and practice mostly focus on the optimization process of obtaining the Pareto optimal solution set, but ignore the screening of the final building optimization design plan from the Pareto optimal solution set and the mining of the data characteristics of the Pareto optimal solution set; with the development of interdisciplinary development, integration Methods for performance simulation and optimization algorithms are becoming more and more diverse, and correspondingly there are requirements for the operability and functionality of their methods, including ease of operation, suitability of the algorithm, potential mechanism of Pareto optimization solution set, and whether it can be used by users Provide clear and objective design decisions and more.

Performance simulation correlates architectural design elements and performance evaluation indicators; multi-objective optimization uses multi-objective optimization algorithms to explore potential solutions and solve multi-variable multi-objective optimization problems. There are two main types of building green performance optimization design methods that integrate performance simulation and optimization algorithms. One is a building energy-saving optimization design method based on mathematical software, and the other is a building green performance optimization design method based on a parametric design platform.

The building energy-saving optimization design method based on the mathematical software MATLAB is based on the interaction between the mathematical software MATLAB and the energy consumption simulation software (such as EneryPlus\TRNSYS). It uses various optimization algorithms to automatically search for solutions with the lowest energy consumption; while the mathematical software MATLAB uses Programming languages, such as C, C++, Java, etc.; they have higher requirements on programming languages and are more difficult to operate; cross-platform interaction between mathematics software and energy consumption simulation software requires joint simulation interfaces, such as BCVTB, jEPlus and MLE+, etc.; only Associated with energy consumption simulation software, optimized performance is limited to energy saving.

The parametric design platform Rhino integrates functional plug-ins such as geometric modeling, performance simulation, evaluation and optimization; although this method has algorithm plug-ins such as Octopus, the available optimization algorithms are limited such as SPEA-2 and HypE algorithms; the optimization results are presented in data tables and scatter plots show that the optimal solution requires manual selection, that is, the optimal solution is highly subjective and uncertain, and lacks the optimization result analysis function.

Contents of the invention

In view of the technical problems existing in the prior art, the present invention provides a building green performance design optimization method and system to solve the existing building green performance design method, which has large optimization performance limitations, complex operation processes, and requires optimal solutions. Using manual selection involves technical issues with greater uncertainty.

In order to achieve the above objects, the technical solutions adopted by the present invention are:

The invention provides a building green performance design optimization method, which includes the following steps:

Select the evaluation indicators used to evaluate the green performance of the building and determine the independent variable parameters that affect the green performance of the building;

Construct the building model to be optimized and import the building model to be optimized into modeFRONTIER software;

According to the selected evaluation indicators for evaluating the green performance of buildings, use modeFRONTIER software to execute a multi-objective optimization algorithm and obtain the Pareto optimization solution set;

Using cluster analysis method or multi-criteria decision analysis method, the Pareto optimization solution set is screened to obtain the multi-objective optimal solution, that is, the building green performance design optimization result is obtained.

Furthermore, the evaluation indicators used to evaluate the green performance of buildings include energy consumption per unit building area EUI, percentage of people who are dissatisfied with the thermal environment PPD, percentage of all-natural lighting time in the space sDA, and annual light exposure ASE.

Furthermore, the multi-objective optimization algorithm is used to find the optimization process, and the minimum value of the energy consumption per unit building area EUI, the percentage of people dissatisfied with the thermal environment PPD and the annual light exposure ASE are used, and the maximum value of the space's full natural lighting time percentage sDA is used. for optimization goals.

Furthermore, the independent variable parameters that affect the green performance of the building include building orientation, window-to-wall ratio, floor height, standard floor area, aspect ratio, window SHGC, window heat transfer coefficient, exterior wall heat transfer coefficient, air-conditioning heating temperature, air-conditioning refrigeration temperature and lighting power density.

Further, the building model to be optimized includes the three-dimensional model of the building structure, building structure parameters, active equipment control parameters and meteorological data of the area.

Furthermore, the modeFRONTIER software is used to perform the multi-objective optimization algorithm process, and the Latin hypercube sampling method is used for sampling to obtain the initial samples; the modeFRONTIER software's optimization module is used to optimize the initial samples and obtain the Pareto optimization solution set; among them, modeFRONTIER software The pilOPT algorithm is built into the optimization module.

Further, the cluster analysis method is used to screen the Pareto optimization solution set:

The data source is the Pareto optimization solution set, and the constraints are the key design independent variable parameters that affect the green performance of the building. Among them, modeFRONTIER software is used to conduct sensitivity analysis on the independent variable parameters that affect the green performance of the building, and the key design parameters that affect the green performance of the building are obtained. Variable parameters; use the cluster analysis module in modeFRONTIER software to perform cluster analysis on the Pareto optimal solution set; among them, the K-Means algorithm is built into the cluster analysis module of modeFRONTIER software.

Furthermore, the multi-criteria decision analysis method is used to screen the Pareto optimal solution set:

The data source is the Pareto optimization solution set, and the constraints are the minimum value of the energy consumption per unit building area EUI, the percentage of people who are dissatisfied with the thermal environment PPD, and the annual light exposure ASE, and the maximum value of the space's full natural lighting time percentage sDA; use modeFRONTIER The multi-criteria analysis module in the software performs multi-criteria decision-making analysis on the Pareto optimization solution set; among them, the multi-criteria analysis module of modeFRONTIER software has a built-in Linear MCDM algorithm.

The invention also provides a building green performance design optimization system, which includes a variable module, a model module, an optimization module and an analysis module;

The variable module is used to select evaluation indicators for evaluating the green performance of the building and determine the independent variable parameters that affect the green performance of the building;

The model module is used to build the building model to be optimized and import the building model to be optimized into modeFRONTIER software;

The optimization module is used to use the modeFRONTIER software to execute a multi-objective optimization algorithm based on the selected evaluation indicators used to evaluate the green performance of buildings, and obtain the Pareto optimization solution set;

The analysis module is used to use the cluster analysis method or the multi-criteria decision analysis method to screen the Pareto optimization solution set to obtain the multi-objective optimal solution, that is, to obtain the building green performance design optimization results.

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

The invention provides a building green performance design optimization method and system. By determining indicators and independent variable parameters for evaluating building green performance, modeFRONTIER software is used to perform multi-objective optimization to obtain Pareto optimization solution sets; modularization and visualization are used Programming language and simple operation process, which effectively improves the optimization efficiency and the accuracy of the optimal solution set; by performing cluster analysis or multi-criteria decision analysis on the Pareto optimal solution set, the screening of the Pareto optimal solution set is realized, which improves the efficiency of building multiple The decision-making efficiency and accuracy of the target optimization design help designers to make design decisions and improve the objectivity and scientificity of decision-making; the optimization performance of the present invention has small limitations, and the selection of the optimal solution is determined by the user's preference, that is, based on the user's preference It can automatically generate optimal solutions with high accuracy.

Furthermore, the energy consumption per unit building area EUI is used as the energy consumption evaluation index, and the percentage of people dissatisfied with the thermal environment PPD is used as the thermal performance evaluation index. Solar radiation heat gain and visible light affect the indoor light and heat environment, and the natural environment is rationally used to meet comfort needs. At the same time, it does not increase energy consumption; the percentage of natural lighting time in the space sDA and the annual light exposure ASE ensure sufficient natural light and avoid glare problems.

Furthermore, using the modeFRONTIER software optimization process, Latin hypercube sampling can not only satisfy the probability distribution covering the entire parameter space; the pilOPT algorithm can automatically stop the optimization, significantly enhancing the operability of the algorithm and greatly improving the applicability of this method; Optimization The algorithm pilOPT has the characteristics of global and local search, ensuring the uniform distribution and good convergence of the Pareto optimization solution set.

Furthermore, by using cluster analysis method or multi-criteria decision analysis method on the Pareto optimization solution set, the impact of building parameters on green performance is quantified, which facilitates designers to explore the mapping relationship between building elements and green performance, and assists users in making designs. Decision-making and selecting the optimal solution that meets the green performance of the building.

Description of the drawings

Figure 1 is a schematic workflow diagram of the building green performance design optimization method described in the embodiment;

Figure 2 is a schematic diagram of the performance optimization framework in the embodiment;

Figure 3 is a diagram showing the sensitivity analysis results of the ASE affecting the annual light exposure of the space in the embodiment.

Figure 4 is a diagram of the Davies-Bouldin index results used for clustering analysis in the embodiment;

Figure 5 is a diagram of cluster analysis results in the embodiment;

Figure 6 is a diagram showing the ranking results of the Pareto optimization solution set using the multi-criteria decision analysis method in the embodiment.

Detailed ways

In order to make the technical problems, technical solutions and beneficial effects solved by the present invention clearer, the following specific examples will further describe the present invention in detail. It should be understood that the specific embodiments described here are only used to explain the present invention and are not intended to limit the present invention.

The invention provides a building green performance design optimization method, which includes the following steps:

Step 1. Select the evaluation indicators used to evaluate the green performance of the building, and determine the independent variable parameters that affect the green performance of the building; among them, the evaluation indicators used to evaluate the green performance of the building include energy consumption per unit building area EUI, and the percentage of people who are not satisfied with the thermal environment. PPD, the percentage of all-natural lighting time in the space sDA and the annual light exposure ASE; the independent variable parameters that affect the green performance of the building include building orientation, window-to-wall ratio, floor height, standard floor area, aspect ratio, window SHGC, and window heat transfer Coefficient, exterior wall heat transfer coefficient, air-conditioning heating temperature, air-conditioning cooling temperature and lighting power density; among them, window-to-wall ratio includes south-facing window-to-wall ratio, north-facing window-to-wall ratio, west-facing window-to-wall ratio and east-facing window to wall ratio.

Step 2. Construct the building model to be optimized and import the building model to be optimized into modeFRONTIER software; where the building model to be optimized includes the three-dimensional model of the building structure, building structure parameters, active equipment control parameters and meteorological data in the area; where, The three-dimensional model includes the morphological parameters of the building structure (such as aspect ratio), functional partitions and window-to-wall ratio information; the building structure parameters include exterior wall heat transfer coefficient, window heat transfer coefficient, roof heat transfer coefficient, window visible light transmittance and window transmittance. Light rate; active equipment control parameters include heating ventilation and air conditioning type, comprehensive performance coefficient of heating ventilation and air conditioning system, personnel density, regional load, heating design temperature, fresh air volume, cooling design temperature, lighting power density and equipment power density .

Step 3. According to the selected evaluation index for evaluating the green performance of the building, use modeFRONTIER software to execute the multi-objective optimization algorithm to obtain the Pareto optimization solution set; in the present invention, the multi-objective optimization algorithm is used to optimize the process, using unit building area The minimum values of energy consumption EUI, percentage of people dissatisfied with the thermal environment PPD and annual light exposure ASE, and the maximum value of sDA as a percentage of all-natural daylighting time in the space are the optimization goals.

Specifically, the modeFRONTIER software is used to perform the multi-objective optimization algorithm process, and the Latin hypercube sampling method is used for sampling to obtain the initial samples; the modeFRONTIER software's optimization module is used to optimize the initial samples to obtain the Pareto optimization solution set; among them, modeFRONTIER software The pilOPT algorithm is built into the optimization module.

Step 4. Use the cluster analysis method or the multi-criteria decision analysis method to screen the Pareto optimization solution set to obtain the multi-objective optimal solution, that is, the building green performance design optimization result is obtained; the building green performance design optimization result is used as the building green performance optimal solution. Optimized solutions for users to make design decisions.

In the present invention, the process of screening Pareto optimal solution sets using cluster analysis methods:

According to the principle of homogeneity of independent variables within clusters and heterogeneity of dependent variables between clusters, several optimal solutions in the Pareto optimization solution set are divided into multiple groups of disjoint solution clusters; according to the designer's preferences, the optimization goals are preset Weight, filter multiple groups of solution clusters cluster by cluster, obtain the optimal solution of preset preferences, obtain the cluster analysis results, and obtain the multi-objective optimal solution, that is, obtain the building green performance design optimization results, as a building green performance-oriented optimal design results.

Among them, the cluster analysis process specifically includes the following steps:

Use the Pareto optimization solution set as the data source.

Use the sensitivity analysis module of modeFRONTIER software to conduct sensitivity analysis on the independent variable parameters that affect the green performance of the building, and obtain the key design independent variable parameters that affect the green performance of the building.

The key design independent variable parameters and optimization goals that affect the green performance of the building are used as constraints; the cluster analysis module in modeFRONTIER software is used, and the K-Means algorithm is used to perform cluster analysis on the Pareto optimization solution set, and the Pareto optimization solution set is randomly decomposed is a set of disjoint clusters; among them, the K-Means algorithm settings include: the maximum number of iterations, the maximum number of clusters, random generator seeds, and cluster numbering.

Use modeFRONTIER software to create a partitioned clustering model; among them, the K-Means algorithm is preset to reduce the distance within the cluster in each iteration, and the Davidson-Boldin index DBI is used to measure the evaluation index of the quality of the clustering model; the Davidson-Boldin index DBI is The ratio of the intra-cluster variance to the sum of inter-cluster distances. The lower the Davidson-Boldin index DBI, the better the cluster quality of the cluster division.

Select the partitioning clustering model with the lowest Davidson-Boldin index DBI, perform partitioning and clustering analysis on the Pareto optimization solution set, and obtain the clustering data table; and obtain the multi-objective optimal solution from the clustering data table, that is, obtain the building green performance Design optimization results.

Sorting and selecting among multiple alternatives is a relatively common but often difficult task. The multi-criteria decision analysis method refers to the decision-making of choosing among multiple alternatives that are conflicting and incompatible; multi-criteria The decision analysis method is to conduct a structured analysis of the Pareto optimization solution set based on pre-specified decision rules, and draw well-founded suggestions for decision-makers to refer to.

In the present invention, a multi-criteria decision analysis method is used to screen the Pareto optimal solution set;

The data source is the Pareto optimization solution set, and the constraints are the optimization goals; the multi-criteria analysis module in modeFRONTIER software is used to conduct multi-criteria decision-making analysis on the Pareto optimization solution set; among them, the Linear MCDM algorithm is built-in in the multi-criteria analysis module of modeFRONTIER software.

Specifically, it includes the following steps:

Select a data source to create a new multi-criteria decision model.

In this invention, the Pareto optimization solution set is used as the data source; the constraint conditions are the optimization goals; that is, the minimum value of the building energy consumption density EUI, the percentage of people who are not satisfied with the thermal environment PPD and the annual light exposure ASE, and the space has all natural lighting time Percentage of the maximum value of sDA.

Using the multi-criteria analysis module of modeFRONTIER software, the linear multi-criteria decision-making algorithm LinearMCDM is selected for multi-criteria decision-making; the linear multi-criteria decision-making algorithm Linear MCDM calculates the utility function, which is the basis for sorting Pareto optimization solutions; the utility function considers the weight and sets the optimization goal The weight can be integrated into the designer's preference control decision-making process; the algorithm settings include whether to exclude wrong design solutions, priority range and indifference range.

Preset preferences and indifference ranges to run a multi-accuracy decision-making model.

In the present invention, compared with the existing method in which the program automatically weights the mean and variance of the target as the target, modeFRONTIER software can directly operate on the mean and variance of the design target, and has a high degree of flexibility and flexibility in selecting the optimal solution. Operability; taking into account designer preferences when making design decisions, i.e., corresponding to variables such as inputs, outputs, constraints, and goals.

Choose the most reasonable solution from a set of available solutions, that is, choose the optimal solution with the largest evaluation value. Each alternative is evaluated to rank the available alternatives; preferences and indifference margins are reflected in the colors of the ranking plot, with green indicating the best ranking, red indicating the worst ranking, and green the one with the largest evaluation value The optimal design solution has the best target performance.

The invention also provides a building green performance design optimization system, which includes a variable module, a model module, an optimization module and an analysis module; the variable module is used to select evaluation indicators for evaluating the building's green performance and determine the impact on the building's green performance. The independent variable parameters of The multi-objective optimization algorithm obtains the Pareto optimization solution set; the analysis module is used to use the cluster analysis method or the multi-criteria decision analysis method to screen the Pareto optimization solution set to obtain the multi-objective optimal solution, that is, to obtain the building green performance design optimization result.

The building green performance design optimization method and system of the present invention propose a building green performance optimization design method based on modeFRONTIER, which effectively solves a number of difficult trade-off problems in green performance; not only can it improve the accuracy of simulation optimization, but also obtain a set of Balance the multi-objective optimal solution of green performance, and be able to mine the data characteristics of the multi-objective optimal solution, and quantify the mapping relationship between independent variables and dependent variables; usually in order to provide a good indoor light environment and thermal comfort, building energy needs to be consumed; Indicators such as energy consumption EUI, thermal environment dissatisfaction percentage PPD, space all-natural daylighting time percentage sDA, and annual light exposure ASE are conflicting with each other; the present invention uses a multi-criteria decision analysis method and a cluster analysis method, and according to the designer The preference evaluates the Pareto optimal solution set and sorts or clusters several optimal solutions to facilitate designers to make design decisions.

Example

As shown in Figures 1-2, this embodiment provides a building green performance design optimization method and system, which is based on the modeFRONTIER software integrating the building green performance optimization design method of Grasshopper/Ladybug+Honeybee, and follows the design, simulation, optimization, Data mining and decision-making construction design process:

First, analyze the initial design conditions and building performance optimization goals, determine the independent variables and dependent variables, and establish a green performance optimization framework in modeFRONTIER software; call Grasshopper/Ladybug+Honeybee for performance simulation to calculate the energy consumption EUI and thermal environment per unit building area Dependent variables such as the percentage of dissatisfied PPD, the percentage of all-natural lighting time in the space sDA, and the annual light exposure ASE; then, use the optimization module of modeFRONTIER software to conduct sampling to generate initial samples, and use the optimization algorithm to automatically optimize and obtain a set of Pareto Optimize the solution set; finally, perform data mining on the Pareto optimal solution set and assist the designer in making design decisions; the specific steps include the following:

Step 1. On the modeFRONTIER software platform, use modules such as independent variables, sampling, optimization, Grasshopper interface, dependent variables, optimization goals, data post-processing and optimal solutions to build a building green performance optimization framework.

Step 2. Based on existing research and practice, determine the independent variable parameters that affect the building’s energy consumption and light and heat performance. In this embodiment, the independent variable parameters that affect the building’s green performance include building orientation, south-facing window-to-wall ratio, and north-facing window wall. ratio, west-facing window-to-wall ratio, east-facing window-to-wall ratio, building floor height, standard floor area, aspect ratio, window SHGC, window heat transfer coefficient, exterior wall heat transfer coefficient, air-conditioning heating temperature, air-conditioning cooling temperature and lighting power density .

Step 3. Utilizing natural resources is an effective measure to reduce building energy consumption. Building energy consumption and photothermal performance are closely related; therefore, building energy consumption and photothermal performance are used as dependent variables; evaluation indicators used to evaluate building green performance include units The building area energy consumption EUI, the percentage of people who are not satisfied with the thermal environment PPD, the percentage of all-natural lighting time in the space sDA and the annual light exposure ASE; the multi-objective optimization algorithm is used to optimize the process, using the energy consumption per unit building area EUI, thermal environment The minimum value of the percentage of dissatisfied persons PPD and the annual light exposure ASE, and the maximum value of the percentage of all-natural daylighting time in the space sDA are the optimization goals.

Step 4. Use the Grasshopper interface of the modeFRONTIER software platform to automatically call Grasshopper/Lady bug+Honeybee (L+H) to simulate the green performance of the building, and calculate the green performance corresponding to the independent variable parameters, that is, the energy consumption per unit building area EUI and the thermal environment. There are four performance indicators: the percentage of satisfied persons PPD, the percentage of natural lighting time in the space sDA and the annual light exposure AS E.

Step 5. Use the optimization module of modeFRONTIER software to perform algorithm optimization; among which, the Latin hypercube sampling method is used for sampling to obtain the initial sample; use the optimization module of modeFRONTIER software to optimize the initial sample and obtain the Pareto optimization solution set; where , the pilOPT algorithm is built into the optimization module of modeFRONTIER software.

Step 6: Use cluster analysis method or multi-criteria decision analysis method to screen the Pareto optimization solution set to obtain the multi-objective optimal solution, that is, obtain the building green performance design optimization results.

Examples

The following is a detailed description of an office building design example in Xi'an. The specific steps are as follows:

Step 1. Build a building green performance optimization framework on the modeFRONTIER software platform, using modular and visual programming languages. The modules required for building green performance optimization include: independent variable module group, Grasshopper interface module, dependent variable module, optimization target module, and optimization module and complete modules, etc.; the optimization process followed is to input independent variables → run the external program Grasshopper → output dependent variables → input new variables according to the optimization algorithm → use the algorithm to optimize → generate calculation results → extract the Pareto optimization solution set → analyze the optimization results → Choose the optimal solution to make design decisions.

Step 2. Determine the building independent variables that affect the green performance of the building, including: building orientation, south-facing window to wall ratio, north-facing window to wall ratio, east-facing window to wall ratio, west-facing window to wall ratio, floor height, standard floor area, aspect ratio, Window SHGC, window heat transfer coefficient, exterior wall heat transfer coefficient, air conditioning heating temperature, cooling temperature and lighting power density, and set the name, unit and value range of the independent variables respectively, as described in Table 1 below.

Table 1 Table of independent variable parameters that affect the green performance of buildings

Step 3. Enter the Grasshopper file for coupling calculation, and the Grasshopper coupling calculation model can be automatically generated, that is, Grasshopper/Ladybug+Honeybee is automatically called for building green performance simulation; among them, Ladybug+Honeybee is a parametric performance simulation plug-in based on Grasshopper, which calls Energy consumption simulation software EnergyPlus and light performance simulation software Radiance have simulation functions such as coupling light environment, thermal environment, wind environment and energy consumption.

Xi'an belongs to a cold area, with a latitude of 34.3°N and a longitude of 108.9°E; obtain Xi'an's standard meteorological database-epw data from the EnergyPlus official website, and select this data as the meteorological data used in the model of this embodiment; calculate the independent variables in Grasshopper/L+H Green performance indicators corresponding to the parameters.

Step 4. Select four indicators to evaluate the green performance of the building. Taking the total energy consumption EUI, the percentage of people who are dissatisfied with the thermal environment PPD, the percentage of all-natural lighting time in the space sDA and the annual light exposure ASE as dependent variables, use unit building area The minimum values of energy consumption EUI, percentage of people dissatisfied with the thermal environment PPD and annual light exposure ASE, and the maximum value of sDA as a percentage of all-natural daylighting time in the space are the optimization goals.

Step 5. Use the optimization module of modeFRONTIER software to perform algorithm optimization.

S51. Use Latin hypercube sampling to generate initial samples. Latin hypercube sampling can not only satisfy the probability distribution covering the entire parameter space, but the sample size is 5 times the number of independent variables, that is, 70; the initial samples are shown in Table 2 below.

Table 2 Initial sample table

S52, the optimization algorithm pilOPT has the characteristics of global and local search, and can automatically stop the optimization, and use the pilOPT algorithm to perform green performance-oriented optimization.

In this embodiment, the total number of optimizations is 1400, that is, there are 1400 optimization solutions, of which the Pareto optimization solution set is 141. As shown in Table 3 below, Table 3 shows the calculation results of the building performance optimization design method described in this embodiment. Pareto optimization solution set and its index value.

Table 3 The Pareto optimization solution set obtained from the optimization calculation in this embodiment

serial number EUI(kWh/m ² ) PPD(%) sDA(%) ASE(%) 1081 114.5 28.0 26.0 22.1 815 72.8 31.1 31.8 13.9 … … … … …

At the same time, the Grasshopper-based Octopus commonly used in the existing technology is used for performance optimization. The optimization algorithm used by Octopus is NSGA2, replacing the pilOPT algorithm used in step 5 above. Use the NSGA2 algorithm to calculate the Pareto optimal solution set and select the optimal solution from it.

Table 4 Performance optimization results of Grasshopper-based Octopus commonly used in the existing technology

Established methods EUI(kWh/m ² ) PPD(%) sDA(%) ASE(%) Optimal solution 75.2 34.8 49.8 75.1

Step 7: Perform sensitivity analysis on the Pareto optimization solution set to obtain the key design independent variable parameters that affect the green performance of the building: As shown in Figure 3, in this embodiment, the sensitivity analysis is performed based on the ASE that affects the annual light exposure of the space. The sensitivity analysis results are shown in Figure 3; in this embodiment, the key design independent variable parameters that affect the green performance of the building include east-facing window-to-wall ratio, aspect ratio, north-facing window-to-wall ratio, and south-facing window-to-wall ratio. , west-facing window-to-wall ratio, heating temperature and cooling temperature;

Step 8: Perform cluster analysis on the Pareto optimization solution set to assist the user in selecting the multi-objective optimal solution, that is, obtain the building green performance design optimization results; specifically including the following steps:

The fourteen independent variables and four dependent variables in the Pareto optimization solution set are used as data sources.

The key design independent variable parameters and performance targets that affect the green performance of the building are used as the variable settings of the cluster analysis model; the scale function is designed to be random, and the distance type is Euclidean distance.

The K-Means algorithm is used to perform partitioning and clustering analysis on the Pareto optimization solution set. The algorithm settings are: the maximum number of iterations is 5, the maximum number of clusters is 10, the random generator seed is 1, and the cluster number is automatically selected.

Set up the modeFRONTIER software to create a partitioned clustering model, and the K-Means algorithm reduces the intra-cluster distance at each iteration. As shown in Figure 4, compared with the Pareto optimization solution of clustering 1-10 clusters, the Davidson-Boldin index DBI concluded that the optimal number of clusters is 3.

The clusters calculated by cluster analysis assist the user in selecting the optimal solution.

Since the Pareto optimal solution set has good diversity and distribution, cluster analysis was performed on the Pareto optimal solution set, and the Pareto optimal solution set was divided into three clusters. The number of solutions in each cluster was 52, 24 and 65 respectively; design elements The impact on building performance is regular. Cluster analysis is used to construct the mapping relationship between design variables and performance targets. Based on the clustering results, users are given certain regional values of design elements to assist users in flexibly choosing the optimal solution.

As shown in Figure 5, assuming that the user prefers lower energy consumption and thermal comfort dissatisfaction, and a better light environment; that is, cluster 0, the corresponding design parameters within the range of cluster 0 can be selected; the south of the optimal solution of cluster 0 The largest proportion is when the window-to-wall ratio is 50%-60%, the window SHGC is 0.1-0.2, the standard floor area is 1005-1555m ² , the aspect ratio is 4-5, and the east-facing window-to-wall ratio is 30%-60% , the north-facing window-to-wall ratio is 50%-90%, the west-facing window-to-wall ratio is 10%-20%, the window heat transfer coefficient is 1-1.2W/m ² ·K, the exterior wall heat transfer coefficient is 0.33-0.38, the floor height It is 3.3-4m, the cooling temperature is 25-27℃, the heating temperature is 20-21℃, and the lighting power density is 10.5-12W/m ² .

Another embodiment of the present invention uses a multi-criteria decision analysis method to screen the Pareto optimization solution set to obtain the multi-objective optimal solution, that is, obtain the building green performance design optimization results to assist the designer in making design decisions; the specific process is as follows:

The independent variables and dependent variables of the Pareto optimization solution set are used as source data.

The linear multi-criteria decision-making algorithm Linear MCDM is selected to calculate the utility function. The algorithm settings include excluding wrong design solutions, the priority margin is 0.05, and the indifference margin is 0.02.

The energy consumption per unit building area EUI, the percentage of people who are dissatisfied with the thermal environment PPD, the percentage of all-natural lighting time in the space sDA and the annual light exposure ASE are set to have the same weight ratio, that is, each accounts for 25%; the multi-criteria decision analysis method is used to determine The optimized solutions are evaluated and sorted, and the optimal solution with the largest evaluation value is selected; as shown in Figure 5, Figure 5 shows the ranking results of the optimized solutions using the multi-criteria decision analysis method.

In this embodiment, the maximum evaluation value is 0.71, and the corresponding design solution number is 858, as shown in Table 5. From Table 5, it can be seen that the corresponding optimization target energy consumption per unit building area EUI and the percentage of people who are not satisfied with the thermal environment The PPD, the percentage of all-natural lighting time in the space sDA and the annual light exposure ASE are 68.5kWh/m ² , 33.4%, 66.1% and 53.5% respectively.

Table 5 The optimal solution automatically searched using this embodiment

serial number 14 independent variable values EUI(kWh/m ² ) PPD(%) sDA(%) ASE(%) 858 (0, 70, 60, 10, 10, 3.6, 1625, 2.5, 0.2, 1, 0.35, 20, 25, 12) 68.5 33.4 66.1 53.5

Furthermore, comparing the optimization results of the existing methods and the present invention, Table 6 shows the optimal solution searched by the building performance optimization design method of the present invention and the existing building performance optimization design method based on Grasshopper/Octopus; compared with the existing Compared with the optimal solution searched by the method, the optimal solution automatically searched in this embodiment increases the energy saving rate by 8.9%, reduces the PPD by 1.4%, increases the sDA by 16.3%, and reduces the ASE by 21.6%, that is, the optimal solution searched by the optimization method proposed in this embodiment The optimal solution performs better.

Table 6 compares the performance optimization design method of this embodiment and the optimization results of existing methods.

A building green performance design optimization method according to the present invention uses Latin hypercube sampling to improve random sampling to ensure the globality and optimization quality of the sample; the pilOPT algorithm has global and local search characteristics, so the design method of the present invention improves search results. The invention has the advantages of excellent efficiency and high accuracy of optimization solutions; this invention simultaneously considers four performance indicators: energy consumption per unit building area EUI, percentage of people who are not satisfied with the thermal environment PPD, percentage of all-natural lighting time in the space sDA and annual light exposure ASE. The optimization algorithm automatically searches for a set of Pareto optimization solutions; however, it is difficult for users to weigh these optimization solutions. modeFRONTI ER has a multi-criteria decision-making function that evaluates and sorts the Pareto optimization solution sets according to the user's preferences to assist designers in making design decisions. , improve the objectivity and scientificity of decision-making; the present invention uses the pilOPT algorithm to automatically stop optimization, significantly enhances the operability of the algorithm, and greatly improves the applicability of the method; the present invention has a data post-processing function and uses the cluster analysis method The Pareto optimization solution set is quantified and the multi-criteria decision analysis method is used to sort the Pareto optimization solution set to assist users in selecting the optimal solution.

The above embodiment is only one of the ways to realize the technical solution of the present invention. The scope of protection claimed by the present invention is not only limited by this embodiment, but also includes any technical scope disclosed by the present invention. Changes, substitutions and other implementations may be easily imagined by those skilled in the art.

Claims (5)

1. The green performance design optimization method for the building is characterized by comprising the following steps of:

selecting an evaluation index for evaluating the green performance of the building, and determining independent variable parameters affecting the green performance of the building;

building a building model to be optimized, and importing the building model to be optimized into modeFRONTIER software;

according to the selected evaluation index for evaluating the green performance of the building, a multi-objective optimization algorithm is executed by using modeFRONTIER software to obtain a Pareto optimization solution set; executing a multi-objective optimization algorithm process by using modeFRONTIER software, sampling by using a Latin hypercube sampling method, and obtaining an initial sample; optimizing the initial sample by using an optimization module of modeFRONTIER software to obtain a Pareto optimization solution set; wherein, a pilOPT algorithm is built in an optimization module of the modeFRONTIER software;

screening the Pareto optimal solution set by using a cluster analysis method or a multi-criterion decision analysis method to obtain a multi-objective optimal solution, namely obtaining a building green performance design optimization result;

the evaluation indexes for evaluating the green performance of the building comprise the energy consumption EUI of unit building area, the percentage PPD of unsatisfied thermal environment, the percentage sDA of all-natural lighting time of space and the annual light exposure ASE;

and screening the Pareto optimal solution set by adopting a multi-criterion decision analysis method:

the data source is a Pareto optimal solution set, the constraint condition is the minimum value of unit building area energy consumption EUI, percentage of dissatisfaction with thermal environment PPD and annual light exposure ASE, and the maximum value of spatial total natural lighting time percentage sDA; selecting a Linear multi-criterion decision algorithm (Linear MCDM) to carry out multi-criterion decision by utilizing a multi-criterion analysis module of modeFRONTIER software; the Linear MCDM calculation utility function of the Linear multi-criterion decision algorithm is a basis for ordering Pareto optimization solutions; the utility function considers the weight, and the weight of the optimization target is set, so that the utility function can be integrated into a preference control decision process of a designer; the algorithm setting comprises whether to exclude wrong design solutions, priority amplitude and non-difference amplitude; presetting preference and indifferent amplitude, and running a multi-accurate measurement decision model;

and (3) adopting a cluster analysis method to carry out a screening process on the Pareto optimal solution set:

collecting a data source by using a Pareto optimal solution;

the sensitivity analysis module of modeFRONTIER software is utilized to carry out sensitivity analysis on the independent variable parameters influencing the green performance of the building, and key design independent variable parameters influencing the green performance of the building are obtained;

taking key design independent variable parameters and optimization targets which influence the green performance of the building as constraint conditions; utilizing a cluster analysis module in modeFRONTIER software, adopting a K-Means algorithm to perform cluster analysis on the Pareto optimal solution set, and randomly decomposing the Pareto optimal solution set into a group of disjoint clusters; wherein the K-Means algorithm settings include: maximum iteration times, maximum cluster number of clusters, random generator seeds and cluster numbering mode;

creating a partitional clustering model by using modeFRONTIER software; the K-Means algorithm presets the distance in the reduced cluster when each iteration is performed, and the evaluation index of the cluster model quality is measured by using the Dyson Bobber index DBI; davison burg Ding Zhishu DBI is the ratio of the sum of intra-cluster variance and inter-cluster distance, with lower davison burg index DBI leading to better cluster quality for cluster division;

selecting a partitional clustering model of the DBI with the lowest Dyson Babbing index, and performing partitional clustering analysis on the Pareto optimal solution set to obtain a clustering data table; and obtaining a multi-objective optimal solution from the clustering data table to obtain a building green performance design optimization result.

2. The method for optimizing green performance design of building according to claim 1, wherein the optimizing process of the multi-objective optimizing algorithm is performed by using the minimum values of the energy consumption EUI per building area, the percentage PPD of dissatisfied with thermal environment and the annual light exposure ASE, and the maximum value of the percentage sDA of the total natural lighting time of the space as the optimizing objective.

3. The method of claim 1, wherein the independent parameters affecting the green performance of the building include building orientation, window wall ratio, floor height, standard floor area, aspect ratio, window SHGC, window heat transfer coefficient, exterior wall heat transfer coefficient, air conditioning heating temperature, air conditioning cooling temperature, and illumination power density.

4. The method for optimizing green performance design of building according to claim 1, wherein the building model to be optimized comprises a three-dimensional model of a building structure, building structure parameters, active equipment control parameters and meteorological data of an area where the active equipment control parameters are located.

5. The building green performance design optimization system is characterized by comprising a variable module, a model module, an optimizing module and an analyzing module;

the variable module is used for selecting an evaluation index for evaluating the green performance of the building and determining independent variable parameters affecting the green performance of the building;

the model module is used for constructing a building model to be optimized, and importing the building model to be optimized into modeFRONTIER software;

the optimizing module is used for executing a multi-objective optimizing algorithm by utilizing modeFRONTIER software according to the selected evaluation index for evaluating the green performance of the building to obtain a Pareto optimizing solution set; executing a multi-objective optimization algorithm process by using modeFRONTIER software, sampling by using a Latin hypercube sampling method, and obtaining an initial sample; optimizing the initial sample by using an optimization module of modeFRONTIER software to obtain a Pareto optimization solution set; wherein, a pilOPT algorithm is built in an optimization module of the modeFRONTIER software;

the analysis module is used for screening the Pareto optimal solution set by using a cluster analysis method or a multi-criterion decision analysis method to obtain a multi-objective optimal solution, namely obtaining a building green performance design optimization result;

the evaluation indexes for evaluating the green performance of the building comprise the energy consumption EUI of unit building area, the percentage PPD of unsatisfied thermal environment, the percentage sDA of all-natural lighting time of space and the annual light exposure ASE;

and screening the Pareto optimal solution set by adopting a multi-criterion decision analysis method:

the data source is a Pareto optimal solution set, the constraint condition is the minimum value of unit building area energy consumption EUI, percentage of dissatisfaction with thermal environment PPD and annual light exposure ASE, and the maximum value of spatial total natural lighting time percentage sDA; selecting a Linear multi-criterion decision algorithm (Linear MCDM) to carry out multi-criterion decision by utilizing a multi-criterion analysis module of modeFRONTIER software; the Linear MCDM calculation utility function of the Linear multi-criterion decision algorithm is a basis for ordering Pareto optimization solutions; the utility function considers the weight, and the weight of the optimization target is set, so that the utility function can be integrated into a preference control decision process of a designer; the algorithm setting comprises whether to exclude wrong design solutions, priority amplitude and non-difference amplitude; presetting preference and indifferent amplitude, and running a multi-accurate measurement decision model;

and (3) adopting a cluster analysis method to carry out a screening process on the Pareto optimal solution set:

collecting a data source by using a Pareto optimal solution;

the sensitivity analysis module of modeFRONTIER software is utilized to carry out sensitivity analysis on the independent variable parameters influencing the green performance of the building, and key design independent variable parameters influencing the green performance of the building are obtained;

taking key design independent variable parameters and optimization targets which influence the green performance of the building as constraint conditions; utilizing a cluster analysis module in modeFRONTIER software, adopting a K-Means algorithm to perform cluster analysis on the Pareto optimal solution set, and randomly decomposing the Pareto optimal solution set into a group of disjoint clusters; wherein the K-Means algorithm settings include: maximum iteration times, maximum cluster number of clusters, random generator seeds and cluster numbering mode;

creating a partitional clustering model by using modeFRONTIER software; the K-Means algorithm presets the distance in the reduced cluster when each iteration is performed, and the evaluation index of the cluster model quality is measured by using the Dyson Bobber index DBI; davison burg Ding Zhishu DBI is the ratio of the sum of intra-cluster variance and inter-cluster distance, with lower davison burg index DBI leading to better cluster quality for cluster division;

selecting a partitional clustering model of the DBI with the lowest Dyson Babbing index, and performing partitional clustering analysis on the Pareto optimal solution set to obtain a clustering data table; and obtaining a multi-objective optimal solution from the clustering data table to obtain a building green performance design optimization result.