z-logo
open-access-imgOpen Access
Predicting the Required Volume Need for Architectural Work Using Artificial Neural Networks in Hospital Buildings
Author(s) -
Aulia Yudha Prathama
Publication year - 2019
Publication title -
journal of the civil engineering forum
Language(s) - English
Resource type - Journals
eISSN - 2581-1037
pISSN - 2549-5925
DOI - 10.22146/jcef.39772
Subject(s) - artificial neural network , computer science , ceiling (cloud) , architecture , volume (thermodynamics) , work (physics) , industrial engineering , artificial intelligence , engineering , geography , mechanical engineering , structural engineering , physics , archaeology , quantum mechanics
Decision-making in construction design has an important role. The need for estimation tools of planning and project management aspects needs to develop. This paper discussed the benefits of artificial neural network methodology to overcome the problem of estimated the needs of the volume of wall paired, ceiling worked pairing, and ceramic floor pairing for architectural work at the designed stage of the building. The average architecture cost of state building is 29%-51% of total construction value. Data from 15 projects was used for being trained and tested by Artificial Neural Network (ANN) methods with 5 design input variables. The ANN helped to estimate the value of volume requirement on the architectural working of Pratama Hospital building project in remote areas of Indonesia. Those input variables include building area, average column span distance, the height of the building, the shape of the building, and a number of inpatient rooms. From ANN simulation, the best empirical equation of P2V5 modeling was used to predict the need of hospital architecture work volume at conceptual stage with best ANN structure 5-9-3 (5 input variables, 1 hidden layer with 9 neurons and 3 output) with result of estimation accuracy a maximum of 96.40% was reached.

The content you want is available to Zendy users.

Already have an account? Click here to sign in.
Having issues? You can contact us here