NEURAL NETWORK PREDICTION MODEL FOR FIELD INSTRUMENTS IN GAS METERING SYSTEM BASED ON PARTICLE SWARM OPTIMZATION

Accurate measurement of temperature, pressure and volume in a gas metering stations is an important aspect to ensure the reliability of billing process. It is highly crucial to have a high degree of measurement accuracy to ensure correct volume of products to be sold. This will ensure value for the money spent by the customer. One of the main problems faced at gas metering systems is an inaccurate gas measurement and unavailability of actual readings from the measuring devices. This scenario will give impact to the instrument readings to become unreliable and this directly affects the -calculation of energy consumptionTTrevious researcher also proposed a prediction model based on healthy instrument reading. However, when the process is in upset condition, the prediction becomes inaccurate. To address this issue, a Neural Network (ANN) prediction model has been proposed to provide a reliable measurement for gas metering systems