OPTIMA (Operator Controllable Tuning using Intelligent Machine Algoritm)

This project focuses on reducing operator-controllable heat loss in a coal-fired power plant using an AI-based recommendation system called OPTIMA.

Problems

In power plants, operators continuously control important parameters such as steam temperature, steam pressure, oxygen level, and air-fuel ratio. Small deviations from optimal conditions can reduce efficiency, increase coal consumption, and raise operational costs. To address this, I developed a real-time monitoring dashboard that acts like an “AI assistant” for operators by comparing current operating conditions with the plant’s historical optimal baseline.

Approach

The system uses historical operational data from the DCS and PI System, which is processed using Python and machine learning algorithms to predict the optimal operating range and identify parameter deviations in real time. Based on these predictions, OPTIMA provides specific operational recommendations to operators, such as adjusting airflow or steam temperature settings. The dashboard was also integrated with the Intelligent Sootblower system to improve overall boiler efficiency.