<![CDATA[Steam power plants are one of the major electricity generation units in many countries around the world. The thermal efficiency of power plants is primarily dependent on decision making by the operator on real time process parameters. This decision-making process currently utilizes human expertise, in conjunction with static setpoints and operating procedures. However, variability in human operator performance and plant operating conditions often leads to non-optimal heat rate values. The purpose of this paper is to develop a conceptual framework for an artificial intelligence-based operator decision-support system for real-time heat rate optimization, integrating Model-Based Design (MBD) and Design Science Research (DSR) principles. The framework presented in this paper is informed by past high efficiency operational experience and machine learning methodology to describe the necessary steps in generating actionable, explainable recommendations for process parameter adjustments. The conceptual framework presented, which incorporates both predictive capabilities as well as domain expertise, is intended to bridge the gap between the development of predictive models and their eventual deployment as prescriptive operational support systems by providing a high-level blueprint of a system design that is expected to lead to more robust and consistent decision making. The key functional components of the framework include data capture, preprocessing, inference modeling and, ultimately, presentation of recommendations on a human-machine interface. An initial, theoretical appraisal of the proposed framework suggests promising potential for improving operational efficiency, reducing fuel consumption, and lowering emissions, and it is expected to serve as a useful reference for ongoing and future development efforts.]]>
Copyrights © 2025
