Yeshwanth Mekala, a doctoral candidate in materials science and engineering, will defend their dissertation titled “Dynamic Heat Flow, Current Distribution, and Wear Estimation in DC Electric Arc Furnace Bottom Anodes Using Fiber Optic Sensing and Thermal Modeling.” Their advisor, Dr. Ronald O’Malley is a professor in materials science and engineering. The dissertation abstract is provided below.

Direct Current Electric Arc Furnaces (DC-EAFs) represent a critical and growing component of sustainable steel production; however, safe and efficient operation depends on reliable thermal monitoring of two highly vulnerable components — the water-cooled upper shell and the pin-type bottom anode — for which conventional thermocouple-based monitoring systems provide insufficient spatial resolution and are susceptible to electromagnetic interference from high-current arcing. This dissertation presents a systematic research program developing, validating, and industrially deploying distributed fiber-optic sensing technologies for real-time thermal monitoring of both components, culminating in a physics-based inverse modeling framework for predictive bottom anode wear estimation.
For upper shell monitoring, Rayleigh backscattering Optical Frequency Domain Reflectometry (OFDR) and Brillouin Distributed Temperature Sensing (DTS) systems were deployed on a 150-ton DC EAF at Big River Steel, Osceola, Arkansas. Rayleigh OFDR achieved a spatial resolution of 2.3 mm at 1 Hz, with distributed temperature data correlating clearly with furnace operational events including burner activation and scrap charging. Brillouin DTS demonstrated complementary advantages of superior vibration immunity and long sensing range, with laboratory calibration yielding a thermal sensitivity of 1.19 MHz/°C and R² = 0.9988, and a 36-hour industrial trial confirming a data missing rate below 1%.
For bottom anode monitoring, Fiber Bragg Grating (FBG) and Rayleigh backscattering sensors were deployed across multiple anode pins in 150-ton and 165-ton DC EAF campaigns. FBG sensors survived complete campaigns of up to 922 hours, providing quasi-distributed temperature measurements that closely agreed with thermocouple reference data and resolved individual tap-to-tap heat cycles and progressive temperature increases associated with anode wear. Analysis of simultaneous FBG data from 48 instrumented pins revealed Joule heating signatures of 1.1–2.2°C superimposed on the conductive baseline, with central pins carrying approximately 40–60% higher current density than outer pins during slag foaming — providing the first in-situ experimental validation of preferential current channeling patterns previously predicted only by electromagnetic simulation.
A transient one-dimensional finite-difference heat transfer model incorporating temperature-dependent material properties and adaptive boundary conditions was developed and validated against the industrial FBG dataset, achieving a mean RMSE of 8.2°C and mean R² = 0.984 across all sensor positions. An inverse modeling framework using proportional feedback control estimated progressive anode pin wear from 1200 mm to approximately 650 mm over a 1050-hour campaign, consistent with post-campaign industrial observations. Parametric sensitivity analysis identified the hot-end boundary temperature as the dominant source of model uncertainty, establishing a clear direction for future improvement.
Together, the sensing and modeling frameworks developed in this dissertation advance DC-EAF thermal monitoring from a sparse, reactive capability to a distributed, predictive one, providing the essential components of a digital twin for bottom anode systems and establishing a foundation for intelligent, safety-aware process control in next-generation steelmaking operations.