جستجوی مقالات مرتبط با کلیدواژه "شبکه عصبی مصنوعی" در نشریات گروه "هنر و معماری"

One of the critical tasks of archaeological studies is to illustrate in detail the geometric properties of pottery fragments. The illustrations are presented as classification and typological concepts for the potteries of a historical period. They are used to investigate the cultural and social structures of that period. Nowadays, most similar studies are conducted by numerical methods and mathematical modeling. This research examines the feasibility of using these methods and algorithms in the classification and typology of ancient potteries. What is important at first glance is introducing a set of geometric properties of the outer profile of the pottery fragments, which are entered to the algorithms in the form of RTC functions. These functions show unique features for an outer profile which are used in the classification and typology of the fragments. The classification of samples is accomplished based on RTC functions and by using an Artificial Neural Network (ANN). The samples are categorized in bowl, jug and jar classes with nearly 95% accuracy. The designed ANN is a three-layer network with one hidden layer that includes two neurons. Network outputs are numeric codes 0, 1 and 2 which refer to bowl, jug and jar profiles, respectively. In the following, by correcting the erroneous samples and out of a total of 209 pottery designs, 128 (61.2%), 42 (20.1%) and 39 (18.7%) samples were placed in bowl, jug and jar categories, respectively. After classification, the typology of the available samples within each class is also performed by the Neighbor Joining (NJ) algorithm. According to the NJ algorithm, the bowl, jug, and jar samples are distinguished by 83, 22 and 27 types, respectively. The accuracy of the typology method is also evaluated between 95 and 97%.

One of the main problems in the world is global warming, which is caused by the building sector and carbon emissions. Researchers have come to the conclusion that managing and reducing building energy consumptions and promoting sustainable building practices are crucial. Particularly in hot areas, exterior shading devices are one of the most significant and useful criteria for sustainable passive architecture design. They can have an impact on managing the building's energy resources, including thermal load, daylight, and adaptive thermal comfort. The application of the shadings is highly effective when shading design parameters have carefully and accurately been studied and designed because an inefficient shading device can easily increase the thermal load and create glare or darkness at the same time, or while keeping daylight indicators within the standard range, increase energy consumptions. Although many studies have investigated the design parameters of exterior shading devices including dimensions, materials, and the location of installation through optimization methods, it seems that none of the researches have considered the effect of shade quality on building performances. To investigate the quality of shading, through field measurement in a residential building in Shiraz, Iran, a model was simulated and validated and then a novel parametric exterior fixed shading device added to the model was created in grasshopper plugin which was able to produce a variety range of shadings. A large dataset of 13600 samples of the parametric shading was produced by applying the LHS technique, which created an outspread community of shading samples and executed the energy simulation for each sample. Then this large data set was used to train and test an Artificial Neural Network (ANN). This ANN was applied as a fast emulator and the searching space for multi-objective optimization through NSGA_III algorithm. The major goals in the optimization process are considered to be the least cooling and heating load, the minimum annual sun exposure (ASE), the maximum spatial daylight autonomy (sDA), and the minimum useful daylight illumination (UDI). These five functions are considered as independent variables. Finally, using an adaptive thermal comfort model, the Pareto front solutions have been categorized based on the Percent of Time Comfortable (PTC). This classification clearly demonstrates that although the optimized shading devices can keep the daylight standard indicators (ASE, sDA, and UDI) within acceptable ranges and reduce the cooling load from 53% to 73% and the heating load by from 8% to 10%, the values of PTC could vary by as much as 33.3% (i.e. 4 months of the year) to 66.67% (i.e. 8 months of the year). This range in PTC value is significant because, although certain optimal shading devices can maintain the PTC in interior spaces at 33.3%, other cases with the same rate of energy saving and daylight standard indicators can raise the PTC value to 66.67%. Therefore, this paper introduces the PTC in adaptive thermal comfort model as a new metric for evaluating the quality of the shading produced by any shading device types.

Noise pollution is an important factor in determining the quality of life in cities and affects social welfare. In this regard, vegetation and green space have an effective role in controlling and reducing urban pollution. Evaluation and modeling of the effective structure and composition of vegetation to control noise pollution in cities reveals the limitation in studies so it has been chosen as the main goal of this research. Noise intensity sampling was performed at 100 stations in parks and passages in urban districts 1 and 2 of Tehran. In order to model the amount of noise reduction (Leq) in the vegetation acoustic wall, artificial neural network modeling was performed using 9 vegetation variables. According to the results, the model with 9-24-2 structure (9 input variables, 24 neurons in the hidden layer and one output variable) with respect to the highest value of coefficient of determination in the three categories of training data, validation and test equal to 0.98, 0.92 0 and 0.9 reveals the best structure optimization performance. Based on the results of model sensitivity analysis, wall width, the mean height of trees, and the mean diameter of canopy with the coefficients of 0.72, 0.44 and 0.15, respectively, were the most effective in reducing the noise intensity in the plant acoustic walls, respectively. The model presented in this study is known as a decision support system in design of vegetation acoustic walls in cities and enables the prediction of the efficiency of these walls with respect to structural variables.

This research is to study the capabilities of artificial neural network (ANN) for predicting solar radiance in an urban context in order to materialize the concept of high-performance architecture. Literature review of the research implies that artificial intelligence (AI) is going to be a new emerging tool to contribute to high-performance architecture, as well as a way of thinking for a significant paradigm shift in environmental sustainability. In the first step, a rule-based method is applied to generate the dataset. In the next step, three different ANN models with different architectures are defined. These models are trained with the generated dataset and regarding the defined algorithm architecture, different results are predicted. The results indicate the precision of each model in predicting the amount of received solar radiation in a new sample location. Finally, these results are compared and the best ANN architecture is selected. The proposed model in this research could be generalized to other similar simulations and it has two main applications. First, it could predict the target parameter instantly without intensive computation. Secondly, it could fit a function for simulating a sustainable parameter only with the given input and output dataset and without needing to know any specific rules for the simulation. The results conclude that AI might be introduced as a comprehensive methodology for sustainable design in contemporary architecture. However, the research shows capability of ANN for outlining and predicting environmental sustainable parameters particularly.

Size of classroom’s windows has significant effects on both comfort level of users and electricity consumption for lighting. Moreover, windows are the main source of energy loss in classrooms in both cooling and heating sectors.
Considering the large number of educational buildings and long life cycle of such them, choosing proper window size is crucial for energy saving in sustainable architectural design. Despite the role that windows have in energy consumption, the literatures are surprisingly limited in providing detailed recommendations for architects in determining the appropriate window size in different climates. Therefore, energy based window design has always been complicated for architects due to the number of involved different components and variables. In order to help the architectural designers, in this paper a new methodology is developed using a well-known artificial intelligence technique. In the proposed methodology, a predictive model for energy consumption cost in terms of window to wall ratio (WWR) and the window facing was created using Artificial Neural Network (ANN). The methodology consisted of a limited sets of direct numerical energy simulations for any specific climatic zone to generate the data required for training the ANN. The DOE-2is suggested in the proposed methodology for direct numerical energy simulations of the daylighting scenarios required for training the ANN. The DOE-2 is a popular and powerful computational model developed with financial support of U.S. department of energy. The trained ANN-based model provides a fast and convenient way of comparing the different daylighting scenarios in designing stage. Indeed, further calculations for direct energy simulations are not necessary and an architect can readily utilize the trained ANN-based model as a powerful tool for forecasting the total energy consumption cost. In order to show the applicability and performance of the proposed approach, 288 daylighting scenarios for a standard classroom in a warm and dry climate, Shiraz-Iran, were simulated to determine the corresponding electric and gas consumption. A square classroom of side 7.4 m is the standard classroom defined by Iranian Organization for Renovating, Developing and Equipping Schools. The DOE-2 is utilized for simulating the defined standard classroom in the study area for estimating the annual gas and electric consumption of the generated scenarios over a 50 years period. Included daylighting scenarios were randomly split into train and test sets. In this study, around 80 percent of data were used for training, and the rest were used to evaluate the performance of the trained ANN. The best training and learning functions for different number of layers and neurons was determined in a trial-error process. Correlation Coefficient (CC), Mean square error (MSE) and Root mean square error (RMSE) are the statistical indices used for training procedure. The best results were obtained with 2 hidden layers and 6 neurons per layer. The 'Levenverg-Marquardt back propagation (trainlm)' and 'perceptron weight and bias learning function (learnp)' were the best training functions found for this research. The results show that the trained ANN can accurately predict the total energy consumption cost (RMSE=0.0811, MSE=0.0066, and CC=0.9672).

Water supply and its management for cities have been the main concerns of urban mangers and planners. The recognition of the amount of water demand and its related effective factors are treated as the important parameters in managing and controlling of urban water consumption. In this study, an expert model for predicting annual water demand using Artificial Neural Network (ANN) has been developed for water supply of the City of Ilam. The input parameters of the model as predictive variables are: Annual Income(X1), Consumption Zone(X2), Area(X3), building area(X4), Family size(X5), Number of valves(X6), and Annual water price(X7). The output parameter is Annual Urban Residential Water Demand (AURWD). By using the collected data and their preprocessing, an optimum structure of the expert model was derived. This was composed of: 3 layers with 7 neurons in the internal level, 10 neurons in the hidden level; and one node as the external layer in which Tansing activating function has been utilized. The comparisons of quantitative and qualitative results of the expert model with the observed quantities based on statistical criteria R2, RMSE and graphical tests indicate that ANN has been effectively implicated in AURWD prediction. Also by using the effective factors, a Multi Linear Regression (MLR) equation for AURWD has been developed. Comparisons of AURWD-MLR equation results with the expert model of AURWD-ANN show the superiority of expert system based on ANN to MLR equation. Therefore, the developed expert model in this study, AURWD-ANN, can be used as an expert decision support system by urban managers and planners.