Fuzzy supervisory controller tuning and analysis for optimal steam distribution in industry

Abstract

Distribution of high-pressure steam in industry must be adequately synchronised to production cycles and their dynamic consumption demands. In many cases high-pressure steam plants have also steam turbine installation, which demands stable input/output operating conditions to produce and maintain nominal electrical energy production. The paper presents an offline concept of Fuzzy supervisory control scheme development, tuning and implementation, in order to avoid unnecessary interference of production cycles. Fuzzy rule database is developed upon process measurements and verified by data-driven process model (auto-associative neural network), and implemented into PLC (commercial platform that supports Fuzzy implementation). Initial set up of Fuzzy input/output scaling factors (gains) is defined by cooperative global optimization algorithm, to avoid implementation issues. The use of Fuzzy supervisory controller resulted in more stable and increased production of electrical energy, where efficiency of the steam turbine output has been raised for approximately 5%.