Prediction of Surface Distress Using Neural Networks
نویسندگان
چکیده
Road infrastructures contribute to a healthy economy throughout a sustainable distribution of goods and services. A road network requires timely programmed maintenance treatments in order to keep roads assets in good condition, providing maximum safety for road users under a cost-effective approach. Surface Distress is the key element to identify road condition and may be generated from many different factors. In this paper, a new approach is aimed to predict Surface Distress Index (SDI) values following a data-driven approach. Later this model will be accordingly applied by using data obtained from the Integrated Road Management System (IRMS) database. Artificial Neural Networks (ANNs)are used to predict SDI index using input variables related to the surface of distress, i.e., crack area and width, pothole, rutting, patching and depression. The achieved results show that ANN is able to predict SDI with high correlation factor (R = 0.996%). Moreover, a sensitivity analysis was applied to the ANN model, revealing the influence of the most relevant input parameters for SDI prediction, namely rutting (59.8%), crack width (29.9%) and crack area (5.0%), patching (3.0%), pothole (1.7%) and depression (0.3%).
منابع مشابه
The Prediction of Surface Tension of Ternary Mixtures at Different Temperatures Using Artificial Neural Networks
In this work, artificial neural network (ANN) has been employed to propose a practical model for predicting the surface tension of multi-component mixtures. In order to develop a reliable model based on the ANN, a comprehensive experimental data set including 15 ternary liquid mixtures at different temperatures was employed. These systems consist of 777 data points generally containing hydrocar...
متن کاملOn the use of back propagation and radial basis function neural networks in surface roughness prediction
Various artificial neural networks types are examined and compared for the prediction of surface roughness in manufacturing technology. The aim of the study is to evaluate different kinds of neural networks and observe their performance and applicability on the same problem. More specifically, feed-forward artificial neural networks are trained with three different back propagation algorithms, ...
متن کاملProduct Yields Prediction of Tehran Refinery Hydrocracking Unit Using Artificial Neural Networks
متن کامل
The Quantification of Uncertainties in Production Prediction Using Integrated Statistical and Neural Network Approaches: An Iranian Gas Field Case Study
Uncertainty in production prediction has been subject to numerous investigations. Geological and reservoir engineering data comprise a huge number of data entries to the simulation models. Thus, uncertainty of these data can largely affect the reliability of the simulation model. Due to these reasons, it is worthy to present the desired quantity with a probability distribution instead of a sing...
متن کامل