The final purpose of this study is developing a soiling simulation system to predict visual deterioration of building outer walls. The simulation system outputs soiling predictive images in the designing process. The simulation system will be established as a helpful tool for designers. If designers use the tool, the cleaning cost of the designed building will be smaller.
Soiling of outer walls is classified into several types by focusing on the building component which soiling appears on. Among them, the type of striped soiling under windows is the most popular and most discomfort. Striped soiling is caused by dust and raindrops. Therefore, the purpose of this primitive report is to propose the method of drawing basic striped soiling caused by raindrops on the designed elevations. The applicable scope of the primitive simulation system is limited to the walls of reinforced concrete buildings which are finished with spraying coating, ceramic tile, and no finishing material. When a wall has an enough eave over the window, or when a window has an enough water drip, the striped soiling caused by raindrops will not appear on the wall under the window.
First research work was collecting photos of existing soiled walls and its building data. By applying a color calibration process, the collected soiling photos were changed to the pixel data that mean soiling quantity. Second work was proposing a model formula. The proposed model formula has four kinds of coefficients of soiling stripes (number, density, center, width). Third work was applying the model formula to all the collected soiling photos. In each soiling photo, the model formula was approximated from soling pixel data.
If every coefficient of the model formula is predicted by building data, soiling will be drawn on designed elevations. Thus, the authors thought about predicting the coefficients from building data. In the collected data, the combination of finishing materials and window frame structures affected the number of soiling stripes. Thus, the soiling data were classified according to the combination. In each classification, the coefficients of the model formula were statistically predicted. The predicted coefficients were substituted into the model formula (soiling predictive model).
By using the soiling predictive model, a piece of software which can output soiling predictive images was made. To confirm the results, the soiling predictive images were compared with photos of already soiled walls. The soiling predictive images are similar to photos of already soiled walls concerning the degree of discomfort. After this report, the authors are going to collect more data, improve accuracy of the simulation, and enlarge the applicable scope of the simulation.
各種検索
サポート
ヘルプ