Keywords

1 Introduction

With the advancement of the times, people are increasingly pursuing the goal of automation. China’s construction industry has developed vigorously in recent years. With the improvement of the national economic level and the acceleration of urbanization, the problems of urban population growth and shortage of industrial land have attracted widespread attention. In order to alleviate the problem of resource shortage, people began to pay attention to the construction of new buildings, and through the application of intelligent algorithms to ensure safety and stability, try to save land costs, reduce the cost of materials required for one-time construction, rationally arrange the construction process and reduce the project cost. The modular components in prefabricated buildings require precise design and manufacturing to ensure their perfect fit and stability during on-site assembly. Artificial intelligence can extract valuable laws and knowledge by learning a large amount of existing design data and building information, and generate more optimized prefabricated component design solutions. By applying artificial intelligence algorithms, quality can be effectively controlled and errors and waste in production can be reduced.

In research in the field of artificial intelligence, domestic and foreign experts and scholars have conducted various literature reviews from different perspectives. Some scholars have proposed methods of using artificial neural networks to solve nonlinear function optimization problems [1, 2]. Some scholars have also conducted related research on genetic algorithms and their applications, and conducted modeling and analysis of buildings in a certain area based on multi-user collaborative work model algorithms. By conducting simulation calculations on different types of buildings during their respective life cycles, these scholars discovered the differences between the results and empirical values [3, 4]. Therefore, this article conducts design research on prefabricated building models based on artificial intelligence algorithms.

With the rapid development of computer technology, artificial intelligence has also received more and more attention in the field of construction engineering. This article will explore the application of intelligent algorithms in the construction of prefabricated building models. This article first introduces the basic theories based on genetic algorithm and ant colony optimization search, and then analyzes the problems of low solving efficiency, large time overhead and slow convergence speed of the traditional manual scheduling method. Finally, this article proposes a new multi-objective optimization solution to solve this problem, and gives simulation experiment results of relevant parameters to achieve efficient calculation of artificial intelligence.

2 Discussion on Prefabricated Building Models Based on Artificial Intelligence Algorithms

2.1 Intelligent Building Model Construction Method

In terms of model construction, iterative methods, evolutionary algorithms or random search methods are generally used. These methods solve optimization problems by applying theories such as parallelization and dynamic programming. They mainly solve the situation where the group trait structure leads to uneven competition among individuals and the characteristics of "small degree of freedom". At the same time, they also take into account the possible occurrence of local optimal solutions and global minimum solutions when the current population cooperates to improve solution efficiency and accuracy. These methods use natural genes or genetic knowledge in the biological world to carry out optimization work such as individual selection and population size evaluation [5, 6]. At the same time, they also draw on the survival of the fittest mechanism of natural species to seek the optimal solution or global optimum, thus becoming a basic idea of intelligent building models, and combined with computer science to form a new round of complex system technology revolution. During the modeling process, this article needs to ensure that the object being processed contains as many features as possible. Improper or wrong selection of parameters will have a greater impact on the final result. Therefore, before performing iterative operations, it is necessary to determine whether the solution area and boundary conditions are appropriate to improve problem-solving efficiency. These methods can perform calculation processing without the need for mathematical models or any input information, without the need to establish precise engineering functional relationships or rule equations, and have good adaptability and robustness to a large number of data samples. However, their disadvantages are that the structural parameters are difficult to select, the accuracy is low, the convergence speed is slow, and they lack the characteristics of a global optimal solution.

Fig. 1.
figure 1

Prefabricated building classification

Full size image

When designing a structure, many factors must be considered, including the interaction between components and possible mutual coupling relationships. According to the classification of prefabricated buildings in Fig. 1, modular buildings account for 0.34; steel structure buildings account for 0.31; light steel keel buildings account for 0.2; prefabricated concrete buildings account for 0.1; 3D printing structures account for 0.05. Typically, this article divides components into multiple units, each unit having independent and unique functions. During the construction process, this article should fully understand what each component represents, not just a product or model. The main basis of this algorithm is the interaction between components and the size parameters of each part, through which the relative positional relationship and mutual coordination degree between each unit are determined [7, 8]. During the modeling process, computers can be used to classify different components to achieve optimization of components and overall spatial arrangement. This article will select the most appropriate materials based on the environmental characteristics of the building and convert them into smart structural models or modular structures. At the same time, it is also necessary to consider that there is a certain connection between the internal systems and subsystems of the building and the external frame.

2.2 Prefabricated Building Model Parameter Selection

The model parameters of prefabricated buildings include a series of elements, such as structural dimensions, wall thickness, and floor slabs. Among these parameters, structure and enclosure have the greatest impact on modeling. When the building has a large stiffness and is subject to height restrictions, the frame supports need to bear horizontal loads. At this time, the cross-sectional strength can be improved by using the side-shift filling method. When the beam is relatively small, selecting external columns as supports can reduce the computational complexity [9, 10]. However, due to the limited height of the building, displacement and deformation may not occur. Therefore, it can consider first designing the frame structure into an equal-length triangle with control functions. Through calculation, the required length, width and other data of each component can be obtained. This data is then analyzed and compared with actual conditions to determine compliance with expectations and specification standards, and is used as a reference for selecting section sizes and determining the height of each beam.

In a frame structure, beams and columns are a relatively independent but interconnected overall conceptual system. When the nodes are closely connected, each reinforced concrete skeleton can be considered to be on the same straight line. On the contrary, if the supports at both ends of a node are unequal or have very different lengths, the node can be considered to be a continuously changing, meandering linear line with little curvature. In actual engineering, various factors need to be considered that have a greater impact on model parameter selection and calculation methods. This model can quickly set parameters and realize functions such as automatic adjustment control [11, 12]. At the same time, the relationship and coordination between the various components in the design can also be adjusted according to the preset parameters to meet the needs of building energy conservation, environmental protection and green ecological sustainable development, and provide users with a more efficient, economical, practical, convenient and comfortable high-quality and reliable structural residential space environment experience.

Various uncertain factors may occur during the assembly process, affecting model accuracy and running time, causing the actual results to deviate from expectations or fail to meet design requirements. Therefore, when conducting experiments in a stopped working state, the interference of these external environmental variables on the final simulation results needs to be taken into account. At the same time, when setting the initialization conditions, the changes caused by these external parameters must also be fully considered. During the modeling process, adjustments can be made based on the required functions of the building at different heights. The plate nodes are fixed together by bolts, connecting the beams and the floor to form a whole, thus achieving a stable load-bearing effect. However, due to the uneven distribution of steel bars, its stiffness, strength and other indicators need to be strictly controlled to avoid affecting the structural safety and architectural aesthetics [13, 14].

3 Experimental Process of Prefabricated Building Model Using Artificial Intelligence Algorithm

3.1 Prefabricated Building Structure

Fig. 2.
figure 2

Prefabricated building models

Full size image

Prefabricated construction is an efficient and fast construction method that uses prefabricated components and on-site assembly and has many advantages. In prefabricated buildings, the main components include beam-column joints, wall joints and plate-rib connections. The beam supports are connected to the main bars through anchor supports, and anchors are set on the plate ribs to increase stability. The walls are connected with horizontal welds and vertically arranged welded panels. The steel skeleton in the frame structure is set on each axis according to different parts. When installing other components, it can position the nodes first and then hoist other parts [15, 16]. However, too much attention should not be paid to the location of main beams, shear centerlines, and other nodes. When considering the installation sequence, it is necessary to coordinate the movement patterns of different parts, forms and spatial conditions. At the same time, it is also necessary to comprehensively consider the distance changes between each node and components or beams to ensure the stable performance and safety of the overall building. Therefore, when designing the prefabricated building model (as shown in Fig. 2), the relationship between the structural functional requirements and various combination methods should be fully analyzed, and reasonable choices should be made. In order to improve calculation efficiency, artificial intelligence algorithms can be used to perform performance testing on prefabricated building models. The structural optimization algorithm can fit the overall solution space with a small number of calculation samples, thereby replacing the traditional calculation model, greatly improving the calculation efficiency, and achieving fast optimization calculations. Such artificial intelligence models can be optimized in terms of underlying computational efficiency.

3.2 Artificial Intelligence Algorithm

Artificial intelligence algorithms can classify, store, transmit and manage input, process and output information in a prescribed manner to realize the machine’s ability to analyze and reason about the received data information. The computing control unit is responsible for execution, operation and maintenance tasks, thereby providing accurate and reliable decision support for operators and making the production process safer, orderly and efficient [17, 18]. Starting from a given direction as the initial value, the optimal solution is obtained based on the known data, and the result is stored in the space. If there are multiple discontinuous distribution points in the local range, it will lead to global performance degradation or system response speed. On the contrary, if there is a limited number of nodes in a graph topology based on random boundary conditions, corresponding processing methods also need to be adopted. In this case, multi-hop subnetworks can be used as a basic model to deal with multi-hop problems and provide support for solving problems. For decision variable \(X = \left( {x_1 ,x_2 ,...,x_n } \right)\), the following constraints are satisfied:

$$ g_i \left( X \right) \ge 0\left( {i = 1,2,...,k} \right) $$
(1)
$$ h_j \left( X \right) = 0\left( {i = 1,2,...,l} \right) $$
(2)
$$ X \in R^n $$
(3)

Assuming there are r optimization objectives in total, and these objectives are different, the overall objective function can be expressed as follows:

$$ f\left( X \right) = \left( {f_1 \left( X \right),f_2 \left( X \right),...,f_r \left( X \right)} \right) $$
(4)

In the formula, f is the r-dimensional target vector, \(g_i \left( X \right)\) represents the i-th inequality constraint, and \(h_j \left( X \right)\) represents the j-th equality constraint. Selecting the relationship between different positions of the object or target in the image according to the required detection function. By determining some connection between the object to be tested and other elements, the eigenvalue distribution state equation in the solution process is calculated. Using these relational expressions for calculation and analysis to obtain the optimal solution (parameters) as the initial concentration or optimization standard. The global optimal solution or approximate solution can also be obtained through iterative search, thereby realizing functions such as the design and organization of the entire algorithm architecture [19, 20].

3.3 Prefabricated Building Model Performance Test

By measuring the distance between the on-site building and components and analyzing the changes between each component, the pressure it needs to withstand in the actual state can be determined. Based on the obtained load conditions, establishing the corresponding relational equations and solve the corresponding formulas. Using numerical calculation methods, the calculation results are stored in the database for later comparison and analysis of optimization processing results. The components are then inspected to ensure they are functioning properly. If anomalies still exist, re-running the test to ensure the equipment is operating within normal values and to verify the new design parameters. Measuring structural parameters and main components, and input the measurement data into the system for calculation. In this process, state vectors are used as basic features to represent. Structural parameters, relationships between key components, and dimensions of key parts and other information at different stages were selected for comparative analysis. Determining the pros and cons of the configuration solutions required in the two situations, and then make the optimal judgment. If the priority is higher and the matching degree is high, it is assigned to the assembly model.

4 Experimental Analysis of Prefabricated Building Models Using Artificial Intelligence Algorithms

Compared with traditional algorithms, intelligent algorithms have unique advantages in efficiency testing. Traditional algorithms require manual adjustment and must maintain the current position during each iteration. Intelligent algorithms achieve fast and accurate data processing through the high automation, high precision and fast calculation characteristics of artificial intelligence technology. This makes many users favor smart algorithms. Because its operation time is long and difficult to control, there is a relative error between the efficiency test results of traditional algorithms and intelligent algorithms. If there is a certain relationship between the system state variable and the expected value or exceeds the set threshold range, it may lead to erroneous identification results, and vice versa. Secondly, when the calculation accuracy of the model is lower than a certain limit, the system may enter an uncertain stage. These factors all have an impact on the performance of intelligent algorithms.

Table 1. Comparison of the efficiency test between the traditional algorithm and the artificial intelligence algorithm
Full size table

In order to solve the problem of efficiency testing of intelligent algorithms, this article needs to seek more accurate testing methods. For example, new testing strategies can be explored to improve testing accuracy and controllability. The test results are shown in Table 1. This article tests the comparative performance of the artificial intelligence algorithm and the traditional algorithm to confirm that the artificial intelligence algorithm has a better effect in optimizing the design of the prefabricated building model. In terms of running speed, the performance result of the artificial intelligence algorithm is 43 m/s, while the performance result of the traditional algorithm is 24 m/s; in terms of operating efficiency, the performance result of the artificial intelligence algorithm is 95%, while the performance result of the traditional algorithm is 74%; in terms of visualization level, the performance results of artificial intelligence algorithms are high, while the performance results of traditional algorithms are low.

This article also strengthens the correlation analysis between system status and expected value to reduce the possibility of incorrect identification. At the same time, attention should also be paid to improving the calculation accuracy of the model to avoid the system entering an uncertain stage. Through continuous optimization and improvement, the efficiency test of intelligent algorithms will be more accurate and reliable, providing users with better services.

Fig. 3.
figure 3

Reliability and robustness testing

Full size image

Due to the limited computing process and processing capabilities of artificial intelligence algorithms, a large number of training sample sets are required when using artificial neural networks as controllers. In addition, the amount of test data is huge and the repetition rate is high. Based on these advantages, this article can improve the reliability index by adjusting the weights, thereby reducing the system error probability and improving the instability and low overall performance of each subsystem in the intelligent building model. At the same time, this approach can also reduce maintenance costs and improve the ability to handle possible future emergencies, thereby reducing potential risks. According to the data in Fig. 3, it can be seen that in terms of reliability, the performance result of the artificial intelligence algorithm is 0.53, while the performance result of the traditional algorithm is 0.43; in terms of robustness, the performance result of the artificial intelligence algorithm is 0.74, while the performance result of the traditional algorithm is 0.67. Through such testing methods, this article can effectively ensure the effectiveness and stability level of artificial intelligence algorithms in actual engineering applications.

Fig. 4.
figure 4

The accuracy of the algorithm

Full size image

This article discusses the issue of artificial intelligence algorithms in the construction of prefabricated building models, aiming to perform accurate calculations within a specific time period. This article uses software to model the structural and dimensional parameter information of the building. In order to represent the distance intervals of different component position value subscripts, this article uses them as feature points, and establishes the optimal solution set of multi-objective optimization problems based on attribute vectors and constraints. It can be observed from Fig. 4 that the accuracy of the artificial intelligence algorithm is 84%, while the accuracy of the traditional algorithm is 65%. After obtaining the optimization results that meet the requirements, this article will select a method with higher accuracy and easy to be promoted and applied in engineering practice to achieve satisfactory results.

5 Conclusion

With the rapid development of artificial intelligence technology, this article has been able to expand human thinking and action to broader, more advanced, and more complex fields. After studying the current status of intelligent algorithms and their applications, this paper proposes a method for constructing assembly models based on distributed computing environments. This article first analyzes the problems that artificial intelligence algorithms and traditional optimization methods may encounter when solving nonlinear programming problems, such as excessive iterations and other defects. Subsequently, this article summarizes the basic principles and mathematical derivation of artificial intelligence algorithms, and verifies their effectiveness and practicality through examples. The research results of this article indicate that the application of artificial intelligence in prefabricated building models can simulate structural behavior, predict risks, and propose safety improvement measures. It can help designers optimize the use of building materials and energy during the design phase to achieve better sustainability. The current artificial intelligence algorithms still have limitations in dealing with the complex structure and diverse needs of prefabricated buildings. Prefabricated buildings involve multi-level and multi-component structural configurations, as well as adaptability issues in different environments. Future research can introduce more complex algorithms and models into the construction of prefabricated building models to meet constantly changing needs.