Effectiveness of Error Recursion Technique on Real Time Closed Loop Pressure Process using validated model by NARX- Recurrent Neural Network tool

Abstract

Analysis of process parameter in a plant is very much essential for the operation in process industries. In this article the real time pressure process is taken into account and process parameters are identified as first order plus dead time (FOPDT) transfer function. System identification of the process is done by Nonlinear Autoregressive exogenous (NARX) Recurrent Neural Network and is validated, proposed for controller design. The work aims to development and implementation of fine tuning of closed loop control response, by Error Recursion Reduction Computational (ERRC) method. Initially, optimization technique Particle Swarm Optimization (PSO) is used to find closed loop control parameter settings. The efficacy of the controller is assessed by the basics of time domain and stability analysis. The durability of the controller is approved by exposing it with both servo and regulatory process. Hence the results demonstrate that the proposed method, which gives least time domain specifications than PSO based PID control settings and also reduces the error much faster.