Fuel oil leak detection in power plant with recurrent neural network and execute in programmable logic controller

In this paper, the use of Recurrent Neural Network has designed in MATLAB software with manual formulation for nonlinear system identification in order to detect fuel oil leakage and then the Neural Network structure has converted to a software block for Siemens SIMATIC programmable logic controller that can be used quite practically inside the project. By comparing the actual data that is taken from the real environment with output of the identified system by the neural network, the leak is going to detect in a few seconds.