The impact of thermal insulating materials in heat loss control in smart green buildings using experimental and swarm intelligent analysis.

Abstract

The efficacy of saving energy standards depends on the ability to anticipate the heat loss of buildings. Environmentally friendly materials, also known as eco-friendly or sustainable materials, have a minimal negative impact on the environment throughout their life cycle. These materials are designed to conserve resources, reduce pollution, and promote sustainability. The characteristics of non-stationary and non-linear heat loss through environmentally friendly materials make it challenging to anticipate accurately. At the same time, many of the industry's presently accessible computational models have been created with this in mind; the majority call for powerful computers and time-consuming computations. The artificial neural network (ANN) has been utilized for prediction, and ground-breaking research has shown the viability of this strategy. This research proposes an artificial neural network (ANN) prototype to estimate construction cooling load usage. ANN is integrated with the vortex search algorithm (VS), stochastic fractal search (SFS), and multi-verse optimizer (MVO) models to compare the three models' outcomes and suggest a more accurate strategy. These techniques make a linear mapping among the output and input parameters, often utilized for modeling and regression. The value of the multiple determination coefficient is also determined. The values of the training R2 (coefficient of multiple determination) are 0.9464, 0.99827, and 0.99522 for VS-MLP, SFS-MLP, and MVO-MLP, respectively, with an unknown dataset which is acceptable. The training RMSE amounts for VS-MLP, SFS-MLP, and MVO-MLP are 0.06433, 0.00619, and 0.01028 for the unknown dataset, which is acceptable. According to the MAE values of 0.0082902, 0.0047834, and 0.0076534 in the training phase for VS-MLP, SFS-MLP, and MVO-MLP approaches and the values of testing MAE error of 0.029107, 0.018167, and 0.029212 for VS-MLP, SFS-MLP, and MVO-MLP approaches, respectively, it is obtained that the SFS-MLP has a lower MAE value. The lowest RMSE value and the higher R2 value indicate the favorable accuracy of the SFS-MLP technique.