engleski

Examples of neural network application in estimate of concrete pressure strength

In the field of civil engineering, as opposed of conventional programs for static, dimensioning, plotting, etc., application of expert systems is still small. In fact, expert systems are more presented in all branch human activities for the purpose of substitution of expert knowledge from various areas of human activities. This paper contains a general display of expert systems and neural networks. Application of expert system in testing typical concrete properties was attempted. Two examples were developed. They are examples of analogue versions of neural network. In the first case, concrete pressure strength after 28 days is estimated and in the other case, apart from pressure strength, volume mass in fresh and hardened state of concrete and slump are estimated. The results, obtained in laboratory research are compared to those obtained by application of the expert system.

neural network ; training ; pressure strength

nije evidentirano