Artificial Neural Networks
Artificial Neural Networks is a system that works similar to a fully developed human brain which is able to store and retrieve data in order to solve complex information and gain knowledge with experience. It includes a symbolic method of intelligent calculations along with data processing with the help of soft-computing.
As mentioned earlier, this network can solve difficult task; has the ability to adopt to changing environment; modify the internal structure to get desired result; generates its own rules from learned examples and at the same time it is inexpensive and fast. It has wide range of benefits and has started to apply in civil engineering field as well and is becoming a common topic of research.
Traditional analytical method for determination of complex problem is getting difficult since formulation becomes tedious. For example, prediction of weather changes is not an easy task as it includes large number of unknown factors which is difficult to understand and it only help us to give a probabilistic solution.
On the other hand, development in Information technology is at a rapid rate when compared to the previous generations. Artificial intelligence is slowly capturing every sector since it provides solutions for complex problems in much simple manner.
Applications of Artificial Neural Networks (ANN) in Civil Engineering
Road safety
Creating a network which can identify the end of row of cars in a particular traffic congestion and display a message to the upstream roadway will help us to reduce intake of vehicles on that path, also saves a lot of time for the drivers. These networks once fully operational, can identify road accidents to improve transport system; can analyze entire road network and infrastructure based on nondestructive measurements using Deflectometer.
It also helps to predict the crack propagation on concrete pavement which helps to predict the service life of the path. New generation vehicles are in-build with Auto-pilot mode which when coupled with ANN system will significantly reduce road accidents and improve the efficiency of transportation system.
Construction Technology & Management
Major issues faced in construction industry are lack of model for finding labor productivity; a method for forecast revenue of disputes while construction; lack of proper cost optimization techniques etc.
Connecting Neural networks these various factors of a construction project along with conventional project scheduling methods such as CPM and PERT can resolve these issues and helps to attain balance between time, cost & resources.
Structural Engineering
Studies in this area have attracted strong interest. Researches are conducted about concrete’s behavior by using ANN. This include predicting failures, crack propagation, development of shear strength according to variation in dimensions of the beam etc.
Monitoring of composite structures using damage assessment technique, identifying and localizing possible imperfections of curtain wall systems and monitoring structural health through changes induced by damage are possible with the collaboration of ANN. Modeling non-linear behavior of steel structures, estimation of seismic response of different types of structures and vibration suppression methods can be studied using this network system.
Soil Studies
Issues faced in geotechnical engineering often include uncertainties in design values of indices which has significant impact. In major cases, it requires great knowledge and experience to solve these types of problems. Since Artificial neural networks has the ability to learn from experience, it will have the modelling superiority over traditional methods. Because there is no need to make assumptions about the fundamental rules governing problem under consideration.
Complex engineering problems such as Liquefaction of soil, loading conditions and its potential failures, soil-pile interactions, Foundation studies can be investigated by ANN modelling.
Water Resource Engineering
Mathematical formulation and modelling for analysing water distribution system, identifying water loss through cracks on underground pipe system, study on retrofitting and rehabilitation of water networks are certain application of Neural networks in the field of Hydraulics.
More studies are yet to done on this field including water quality assurance, expansion of sewage network system, methods to increase the efficiency of sewage treatments etc with the help of ANN system.
From studies conducted in past two decades, it shows promising results in using Artificial Neural Networks in Civil Engineering field and their usefulness in complex problems in construction works. However, learning this network system requires longer training times and proper training sets.
Handling large volume data is difficult task since it covers all the satisfactory solutions. Even though, this method can efficiently replace conventional approaches in problem solving.