A Comparative Study of AI-Enabled SCADA and DCS Integration for Smart Grid Optimization in Combined Cycle Power Plants (2018–2026)

Abstract

This study examined the comparative effectiveness of AI-enabled Supervisory Control and Data Acquisition (SCADA), Distributed Control Systems (DCS), and integrated SCADA–DCS frameworks for smart grid optimization in combined cycle power plants over the period 2018–2026. A quantitative longitudinal comparative design was employed using operational datasets from 54 power plants, comprising 2,592 observations across three system configurations. Descriptive and inferential statistical analyses were conducted to evaluate key performance indicators, including response time, control latency, thermal efficiency, system reliability, and fuel consumption. The results revealed statistically significant differences among system types (p < 0.05), with integrated SCADA–DCS systems consistently outperforming standalone configurations. Specifically, integrated systems achieved the lowest average response time (180 ms) and control latency (135 ms), compared to AI-enabled SCADA (245 ms; 190 ms) and AI-enabled DCS (210 ms; 160 ms). Thermal efficiency was highest in integrated systems (58.9%) relative to DCS (55.7%) and SCADA (53.2%), while downtime was reduced to 7.1 hours annually compared to 10.2 hours for DCS and 13.6 hours for SCADA systems. Effect size analysis indicated large impacts for thermal efficiency (η² = 0.45) and response time (η² = 0.42), confirming the practical significance of system integration. Regression analysis further demonstrated that integrated SCADA–DCS configurations were the strongest predictor of performance outcomes (β = 0.62, p = 0.000). The findings highlighted that combining supervisory monitoring with process-level control, enhanced by artificial intelligence, significantly improves operational efficiency, reliability, and adaptability in dynamic smart grid environments.