Thermocouples are the most widely used temperature sensors in industry. However, they lose accuracy over time when used in harsh environments like furnaces. This degrades product quality and causes inefficiency. Replacing them is expensive and time consuming.

Our research aims to improve temperature control in industrial furnaces by making better use of the thermocouples already installed. Specifically, we will study furnaces used to heat treat high value items like aerospace components.

We plan to apply statistical techniques to networks of thermocouples - either multiple wires in a single probe or multiple probes distributed throughout a furnace. By analyzing relative drift between thermocouples, we can better estimate uncertainty and detect sensor degradation.

The goal is to control process temperatures to within 2 °C over a period of 2 months for temperatures above 1100°C. This will reduce overheating and improve product quality without requiring costly sensor replacements.

Improved temperature control through software enhancements also facilitates the transition to more efficient ultra-low temperature furnaces. Our research enables manufacturers to push their equipment further and save on energy bills.