Organic Rankine Cycle Control using Neural Networks and Variable Structure Control
Abstract
This research work proposes an adaptive optimal Lyapunov control for a Trans-critical Organic Rankine Cycle (TORC) to address disturbance rejection problem. TORC is a thermodynamic cycle for converting low temperature heat into electrical power where the system operates with an organic fluid with low boiling temperature at supercritical or trans-critical conditions. This thermodynamic cycle is similar to the traditional steam Rankine cycle, but it is capable of recovering low grade heat into useful power. The control-oriented challenge of utilizing this system at nominal efficiency is that low grade heat resources have a fluctuating nature which does not allow the heat recovery system operates at nominal operating conditions. In this regards, a control-oriented nonlinear model of TORC system is developed, and an Adaptive Variable Structure Control (AVSC) using Cerebellar Model Articulation Controller (CMAC) neural networks is proposed to maintain the system at nominal operating conditions while rejecting the disturbances. The variable structure control is a well known robust control strategy suitable for this application where CMAC neural network method integrates into the control such that the final control is not only robust but also adaptive to the disturbances.
Description
Keywords
Engineering--Mechanical
Citation
Badkoubeh Hezaveh, B. (2016). Organic Rankine Cycle Control using Neural Networks and Variable Structure Control (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/27225