Real Time Fault Detection and Analysis Using Neural Networks for Double Pipe Heat Exchanger

摘要

The heat exchanger is an equipment used to transfer of heat from one medium to another medium. Heat exchanger is widely used in many industrial sectors like petrochemical plants, synthetic plants, thermal power plants and etc. Due to changes in process parameters, the general heat exchange rate will be severely influenced. To increase the efficiency of general heat exchange rate, the conceivable deficiencies created in the heat exchanger must be recognized. For this purpose, there are various fault detection methods are to be considered. Among them, from the process history based methods, Neural Network (NN) can be proposed to detect various possible faults in heat exchanger. This technique requires vast measure of process information. The constant readings from a double pipe heat exchanger are acquired under fault and normal working conditions. By detecting various faults, the performance of the process can be improved.