Multi-objective optimization design of steel structure building energy consumption simulation based on genetic algorithm

1 Introduction

With the development and advancement of China’s building energy conservation, at present, China is conducting energy analysis and energy-saving scheme comparison and selection of buildings. Although China has shifted from traditional energy consumption analysis to building energy consumption analysis based on building information model (BIM) platform, BIMs often lack their true meaning. Orthogonal experiment method is often used for comparison and selection of energy consumption simulation and energy saving schemes. The application process of BIM technology in building energy efficiency design is shown in Figure 1 [1]. Although this method is selected under single factor parameters, ensuring that the building exhausts all its design options, model modification will greatly increase labor, time and cost. When the design plan emphasizes timeliness, the disadvantages of building energy consumption analysis through this method are particularly obvious, it is difficult to reflect the advantages of building energy consumption simulation and design scheme comparison based on BIM [2]. In addition, in the field of China’s building energy analysis and research, information extraction process of building energy consumption analysis based on BIM platform is usually achieved by means of schedules and software development [3]. However, the information that can be extracted through the detailed list is not comprehensive, for example, specific parameters such as the thermal performance of the building cannot be obtained through the schedule. Furthermore, a lot of research works adopt the method of developing data interface, realize data interaction and parameter extraction between building information [4]. Obviously, there are certain industry barriers in the development of data interfaces, and they need to be upgraded with the upgrade of the software, which is more difficult [5]. Because steel structure system has light weight, easy installation, short construction period, good seismic performance, investment recovery, less environmental pollution, less engineering damage or scrap, and steel renewable advantages, steel xu residences are gradually becoming a new force of residential industry. The former chief engineer of ministry of construction comrade Yao Bing has pointed out that the 21st century building structure is the metal world. Therefore, countries all over the world are vigorously developing steel structure. The development and application of steel structure reflects the comprehensive economic strength of a country to a certain extent.

Figure 1

Application process of BIM technology in building energy-saving design.

2 Literature review

In response to this research question, in order to reduce building energy consumption, Nejati et al. conducted energy analysis based on factors affecting energy consumption and data trends of two different commercial buildings, established a benchmark for the energy efficiency of commercial buildings, and realized energy efficiency analysis by multi-layer perceptron (MLP) based on artificial neural network (ANN) [5]. Somervell et al. designed experiments on the weather conditions and types of buildings, and conducted a standardization study of energy consumption, and compared with the design data [6]. El Hraiech et al. proposed a new hybrid forecasting system to predict the power system of the building, which includes four modules: data preprocessing module, optimization module, prediction module and evaluation module [7]. Li et al. aimed at the application of foreign energy consumption software and the status quo of building energy saving, the combination of BIM technology and building energy analysis, and proposed the automatic identification process of building high energy consumption [8]. Ge et al. through the sub-item measurement analysis of building energy consumption, aiming at the lighting energy consumption and equipment energy consumption, established a decision tree model and a linear regression model for building energy consumption analysis [9]. Si et al. fitted the calculation method of building energy efficiency equivalent energy consumption by linear regression method, and the calculation model and threshold range were obtained [10]. Ascione et al. analyzed and calculated the RC thermal network model, while displaying the RC model as an example, the shortcomings of unstable input parameters and single prediction form were proposed [11]. Miri et al. used ENVI-met to simulate the urban microclimate and building energy consumption through EnergyPlus, an analytical method to quantitatively evaluate the impact of microclimate elements on building energy consumption by collaborative simulation [12]. Yun analyzed building energy consumption based on China’s Qinghai-Tibet Plateau, by collecting and analyzing meteorological data such as solar radiation in this area [13]. Su based on the results of building energy consumption simulations, used the Simulink module in MATLAB to model traditional electrical equipment, the model is able to determine and quantify the internal gain, the impact on indoor comfort and energy consumption of buildings [14]. Taflanidis et al. studied the operation of the open air supply system, through the assessment of open air supply, in order to promote the application of mixed ventilation concept in architectural design. After verifying the predicted average voting model (Predicted Mean Vote, PMV), the effects of air supply speed and blade spacing on the flow field and thermal field were studied. Adopt response surface method based on compound design of five factorial center, the optimal parameters of indoor thermal environment comfort are studied [15]. The calculation method of building energy consumption in China is statistically sorted out and summarized. China’s current energy consumption model developed in 2010, by Tsinghua University in 1 year for building energy consumption model calculation simulation cycle, with the characteristics of different types of building energy consumption, in Chinese provinces, municipalities directly under the central government, autonomous regions, special administrative region of 35 provincial administrative units based on statistical analysis of building, China’s energy is divided in to five modules. To ensure the privacy and security, electronic voting system based on homomorphic encryption is presented. The proposed scheme is suitable for multicandidate elections as well as for neutral votes [16]. Lightweight encryption is proposed for the encryption of data which is stored in the cloud. The scheme provides better encryption with less overhead [17]. Randomization technique cannot prevent the recent attacks. Elliptic curve cryptography is another lightweight technique used to implement security and privacy of the system. Parallel architecture is also proposed for prevention of attacks [18]. Identification-based encryption is used to strengthen the sensitive data confidentiality in public cloud storage. Files are encrypted with the user identity. Statistical design also shows that overhead cost time is insignificant for large file sizes [19]. In communication, we know that mobility is the major concern. Vertical handover can be used by integrating with Multi-criteria Fuzzy-based algorithm for optimal network selection [20]. Blockchain-based secure data management can be used for providing the security and tamper proof data. Blockchain uses distributed ledger technology for doing the transaction safely and securely on the network [21]. Bayesian-based learning algorithm is used to develop the estimated parameters for the model we have used for implementation in the network [22]. Data integrity is another concern which we need to take care in the communication. Forensic tools can also be used to detect a particular forgery [23]. Gradient-boosting technique can also be used for validation of the parameters and blockchain-based trust model used for protecting the integrity of the system [24]. Based on current research, the author proposes a multi-objective optimization design for energy consumption simulation of steel structures based on genetic algorithms (GAs). First we will outline the multi-objective optimization theory and GA, then elaborate on the optimal design theory of steel structure, experiment and discuss. GA and BIM technology are complementary in terms of cycle and cost calculations: combination of GA and BIM digital information simulation platform can provide mathematical basis for BIM schedule management and cost management, the flow chart of the GA optimization process is shown in Figure 2. At the same time, because BIM is an information platform for the entire construction industry, extensive information exchange and information updates, it can provide a large amount of reference data for the calculation of GA. When there is information that affects the energy consumption and cost changes in steel structures, the system can automatically change the GA parameters, re-optimize the calculation and even change the construction plan. Experiments show that the 5D BIM construction period and cost optimization model constructed by the author meets the actual requirements of the project, and can be applied and promoted on a large scale.

Figure 2

Flow chart of genetic algorithm optimization process.

3 Methods

3.1 Multi-objective optimization theory and GA

1. Overview of multi-objective optimization theory

The exploration of optimization problems initially started from a single problem, solving single objective optimization problem, since its inception, has received extensive attention and has been quickly applied in all walks of life [25]. As the economy develops and technology advances, people are beginning to find that single-objective optimization problem solving has become more and more one-sided [26]. In work and life, more problems are not the solution of a single problem, it is the multi-objective solution process under resource constraints [27]. The development of science and technology makes the multi-objective solution of the described problem possible, the multi-objective optimization problem can be described in the following form through mathematical language:

(1) min [ f 1 ( x ) , f 2 ( x ) , f 3 ( x ) , … , f m ( x ) ] s ⋅ t l b ≤ x ≤ u b , A e q ∗ x = b e q , A ∗ x ≤ b ,

where f1(x) is the objective function describing the problem; x is the variable; lb is the lower limit of the value of the variable x; ub is the upper limit of the value of the variable x; Aeq * x = beq is the linear equality constraint of variable x; A * x ≤ b is the linear inequality constraint of variable x.

As can be seen from Figure 3, under resource constraints, an increase in the value of an objective function requires the sacrifice of another objective function value. In other words, it needs to be at the cost of reducing the value of another objective function, the solutions A and B corresponding to the value of the objective function are called Pareto optima.

1. Overview of GA theory

Figure 3

Multi-objective optimization problem.

GA is mainly based on the principles of genetics and the theory of biological evolution [28–32], at present, it has been widely used in various industries to find the optimal solution [33–36]. The process of GA to find the optimal solution is a competitive process [37–40]. Survival of the fittest, those who are not adaptable are gradually eliminated [28]. Individuals who “survive” are so-called excellent genes, individuals that are “eliminated” are the genes that have been filtered out [41]. In a computer platform, the GA embodies the genetic process in the biological world through coding [42], the parameters describing the problem are coded into multiple sets of chromosomes [29]. Coding refers to the application of the computer to the simulation calculation of the optimized solution [43,44], and expression of the solution data in the optimized solution space corresponding to the target as structural data on the chromosome [45,46]. The data can be combined in different ways, multiple individuals (i.e., chromosomes) form a group in the concept of GA [47]. Selection based on GA method, crossover, mutation, a process of multiple iterations, that is, the solution process of the intelligent algorithm [48]. The selection process is the process of selecting excellent chromosomes from the population, by selecting the individuals who are left to enter the next iteration of the process [49], who will have the opportunity to continue to multiply as a parent. Through crossover, a chromosome with the characteristics of the parent chromosome can be obtained. The mutation theory based on genetics, which is generally difficult to occur, in the optimization process, a smaller value is usually selected [50].

3.3 Data modeling

The optimization of steel structure energy consumption needs to consider two stages: The first stage is the project design and planning. At this stage, under the constraint environment set by the owner, simplify the target variable, auxiliary historical data, and solve using algorithms. The second stage is the project construction stage, managers with rich experience in accordance with the contract documents and actual conditions, realize optimization through reasonable arrangement and adjustment of construction plans (such as assembly line arrangement, personnel scheduling, vehicle and mechanical path planning, etc.). But in the construction stage, any accident which affects the energy consumption of steel structure construction may happen, for example, uncontrollable weather changes, uncoordinated construction, fluctuations in market prices, changes in interest rates, etc. Therefore, in order to improve the practicability and universality of the optimization results in the steel structure energy consumption optimization model, the addition of BIM is crucial [34]. In order to shorten the step length of the algorithm and accelerate the growth of the population, this algorithm makes the following assumptions: (1) There are no other resource constraints; (2) There is no rework problem in each process. The relationship between construction period and cost is shown in Figure 4.

Figure 4

The relationship curve between time and cost.

When using GAs to solve optimization problems, we choose a continuous process without time reserve for optimization, freely choose between the processes of each activity, and establish a mathematical model formulas (2)–(5):

(2) C = min ∑ C n ( i , j ) − C e ( i , j ) D n ( i , j ) 2 − D e ( i , j ) 2 ⋅ D ( i , j ) 2 + C e ( i , j ) D n ( i , j ) 2 − C n ( i , j ) D e ( i , j ) 2 D n ( i , j ) 2 − D e ( i , j ) 2 + Δ E C ⋅ D ( i ) ) ,

(3) T = ∑ D i ,

(4) s ⋅ t D e ( i , j ) ≤ D ( i ) ≤ D n ( i , j ) ,

(5) T ≤ T pro .

4 Results and analysis

There is a building plan to be constructed. Through the investigation and analysis of the project contract documents, obtain the energy consumption and cost of its basic engineering part (from the beginning of the precipitation activity to the end of +0) contract plan for normal construction and emergency construction. Perform optimization analysis on this part. If this part of the project is under normal construction, the time required to complete the project is 132 days, and the required cost is 4.724 million yuan. If this part is constructed in an emergency state, the time required for the completion of the project is 81 days, and the required cost is 5.209 million yuan. Using the combined GA and BIM model to optimize, the values of the GA algorithm related parameters are obtained, as shown in Table 1.

Run the program several times and the results show:

1. The energy consumption of steel structures gradually tends to a stable value as the number of iterations increases (Figure 5). After reaching stability, change the number of iterations or other genetic parameters, the value and the combination of activities remain unchanged (Figure 6). For example, when the schedule is compressed to 105 days, the planned cost gradually tended towards and stabilized at 4.3 million yuan, and the combination of the duration of each activity also remains unchanged [60–62].

2. The envelope diagram of energy consumption optimization results for steel structures is shown in Figure 7. Project managers can generate Pareto frontiers based on GAs, under certain constraints of time, or select an optimized set of process combinations with flexible activity space.

3. The collision check BIM based on the energy cost combination of steel structure optimized by GA, automatically generate pipeline construction arrangements, and carry out collision detection on the progress of various activities in the construction arrangement, automatically adjust the construction plan or design drawings according to the inspection results. In order to verify the feasibility of the optimization model based on GA and BIM technology, the error results of GA, GA + BIM and actual construction are compared, respectively. Analyze and compare, the GA + BIM optimization model is better than the GA model in fitting the actual project. Make full use of BIM technology and the advantages of GA, the BIM technology, GA is combined with architectural engineering construction schedule optimization theory, building construction engineering construction schedule based on BIM and GA optimization system, solved the construction schedule planning early quantities such as data acquisition, construction progress, optimization of quantitative analysis and optimization model Solve difficult problems.

Figure 5

Iterative optimization curve.

Figure 6

The best combination of current activities.

Figure 7

Pareto de-envelope diagram.

5 Conclusion

GA and BIM technology are complementary in terms of cycle and cost calculations: combination of GA and BIM digital information simulation platform can provide a mathematical basis for BIM schedule management and cost management. At the same time, because BIM is an information platform for the entire construction industry, extensive information exchange and information updates, it can provide a large amount of reference data for the calculation of GA. When there is information that affects the energy consumption and cost changes in steel structures, the system can automatically change the GA parameters, re-optimize the calculation and even change the construction plan. In a practical case, the application of the construction schedule optimization system and optimization model is carried out to verify its feasibility and effectiveness, so as to provide a basis for project managers to formulate a scientific and reasonable construction schedule plan, and help to improve the application effect of BIM technology in the field of project management and construction schedule management. Building energy-saving design and optimization design of building indoor thermal environment comfort, it is a huge system project integrating thermodynamic system, thermal balance principle, structural design, etc. For other more and more detailed influencing factors that affect building energy consumption and indoor comfort, it has not been considered, factors such as structural design and thermal bridge effect need to be further studied and demonstrated.