Abstract:
The Project describes a design model of sunshade group which is suitable for site and time of usage. This model designs from the production and usage of two aspects of requirements, to ensure that the design meets the practical and economic premise of the final optimization of the design objectives. The parameters can be input according to the specific characteristics of materials and the characteristics of the site to make a suitable design of sunshade group.
Keywords: multi-objective optimization, generative design, parametric modeling, architecture, flexibility
Introduction
The Design Problem
We attempt to design a model that can make a suitable design according to the site setting and the time period of shading in need.
In a park, we assume that there are prefabricated bases for poles in the site, which are distributed on a grid. According to the length of the selected base and poles, they can be combined to form sunshades of different sizes and heights to provide shade for activities that happened in the site. But at the same time, the variability of pole length also means the variability of sunshade shape.
However, due to the different time that activities happened and characteristics of the site in the park, there are more specific requirements for the shadow area generated by the sunshade group. First, because the time and place of activities in the park are different, the sunshade should be adjusted according to the sunshine in a specific period of time to meet the shading requirements of the site in a specified time period. Secondly, due to the limits of materials and structures, the size of the sunshade composed of poles is limited to a certain range, resulting in different sizes and shapes of sunshades and different types of groups.
Methodology
According to our assumption, we divide the generation of the sunshade combination group into four steps: 1. The logic of sunshade generation design; 2. The design of performance metrics; which is used to measure whether each design meets the requirements of the problem; 3. Select the better designs that meet the performance requirements through the model design; 4. According to the situation, in reality, take some non-quantifiable performance into consideration and discuss the practice and the possibility of completion of the ground to make it possible to develop into a final workable design.
Design space model
In order to select the sunshade group that meets the shading requirements on the site, and reflect the various forms as much as possible, we first select 15 points randomly on the site, and these 15 points will become the center of the projection of the sunshade on the site,, and select the base of four poles of each sunshade with these 15 central points, so as to ensure that the vertical projection on the site of each sunshade on the ground is square. The side length of a square is selected from one of the four preset lengths at random. The poles of each sunshade can be adjusted within a certain range (2–5meters), so that they can meet the rationality of pedestrian passage, structure and practical use. Under these conditions, different types of sunshade groups are generated, and their performances are discussed in the following settings.
In this model, for each sunshade:
- The side length of the projection perpendicular to the ground of the sunshade (L)
- 2. Length of sunshade pole (K)
3. Position of each sunshade
Figure 1. The logic of sunshade generation design
Figure 2. The optimization constrain of sunshade design
Performance metrics
In the model design, we introduce a variety of parameters to make each sunshade have a rich shape and make the sunshade group more diverse. In addition, the model needs to include a set of metrics to tell the algorithm ofich design is better. We provide a set of metrics from the requirements of use and requirements of production to determine the optimization objectives and constraints.
- The shadow area produced by sunshade group on the site. This should be maximized. In other words, the light area on the site should be minimized.(objective)
- 2. Minimum overlap area, which is the overlapping area projected by the sunshade on the ground. (constrain)
This set of an objective and a constraint completes the description of the model and allows the genetic algorithm to automatically search for possible design ranges to find a set of effective optimization designs.
Model Optimization
Using this model, we perform optimization with the owing settings in discover:
- Number of designs per g eneration:50
- - Number of generations:50
- - Mutation rate: 0.05
We can visualize the results by plotting them relative to the objectives of the optimization. The figure shows the different design schemes explored in the optimization. The x-axis represents the shadow area, and the y-axis represents the overlapping area of the sunshade, that is, the overlapping area of the projection perpendicular to the site of each sunshade.
In order to meet the elements of the mum shadow area and financial requirements of the mum overlap area. In discover, we have selected five designs that are closer to the optimal performance we required. In the figure, these designs will optimize the selection of designs closer to the optimal performance area in the lower-left corner.
Figure 3.
During the optimization process plotted according to the objective and constrain.
Figure 4.Sunlight analysis of designs selected from the model.
Results
After a series of designs which are close to the optimal solution are selected, people can further analyze these designs. Because this design model can not discuss the elements that can not be quantified, such as the modeling design of each sunshade and the adaptability of the whole design model to the surrounding environment of the site, or whether the sunshade groups can meet other requirements of activities on the site, for example, he sunshade group can cover the main moving lines of people and so on...
The design of this model is quite simplified, only involves the approximate shape, planning and position of the sunshade. Later, due to the mutual coverage of the sunshade, the sunshade poles may sometimes pass through the film. The secondary optimization allows us to refine the design, adjusting the length of the rod and the position of the sunshade, so as to avoid the problem of making the poles through the film in the design The unreasonable shape of the sunshade is caused by the unreasonable length of the poles.
When the final design is selected and the secondary optimization is completed, the sunshade group can be assembled as shown in the figure according to the design.This design meets our design purpose: to get the maximum shadow area in the required time period, and increase the utilization of the materials, so as to minimize the overlap area on the site.
Figure 5. Final design after secondary optimization.
Conclusion
In this paper, we introduce a design model of the hade group. This model can meet the requirements of different sites and time periods of activities, and make good economic design planning. Although there is still the lem of staggered shed in the model design, which needs to be optimized later, this method can be adjusted according to the application of new materials and the requirements of design.
References
Danil Nagy, Dale Zhao, David Benjamin. Nature-Based Hybrid Computational Geometry System for Optimizing Component Structure.
Danil Nagy. Project Discover: An application of generative design for architectural space planning
Keinhard Koenig, Katja Knecht. Comparing two evolutionary algorithm based methods for layout generation: Dense packing versus subdivision. Artificial Intelligence for Engineering Design, Analysis and Manufacturing, (2014),28.
