Neural control design implemented in a PLC for a laboratory plant

摘要

This paper describes how advanced control strategies subroutines of a high-level PLC were implemented, on a laboratory plant with industrial instrumentation. The objective proposed was to implement regulatory flow control throughout its operation range, in spite of its nonlinear behavior. The experiences were oriented to produce learning activities on how to implement advanced control strategies on conventional industrial controllers. First we applied a gain-scheduling PI control strategy, allowing a good control performance regardless of the system operating region. The results obtained here were applied to train an ANN to implement neural control capable of controlling the system in the same way. The results indicate that both techniques can be used effectively for the control of industrial nonlinear plants using their own PLC's.