Electrical capacitance tomography is a non-invasive diagnostic tool widely used to visualize gas-solids flow dynamics in fluidized beds. However, while robust sensors capable of surviving temperatures up to 800 °C have been developed, temperature-induced permittivity variations remain a formidable challenge for accurate phase reconstruction. In a study published in ENG. Chem. Eng., researchers at the Chinese Academy of Sciences and collaborators present a comprehensive numerical and experimental investigation to quantify and mitigate thermal interference in high-temperature ECT measurements.
The team conducted numerical simulations across a broad geometric matrix, featuring sensor outer diameters from 27 to 1067 mm with varying wall thicknesses, to represent laboratory to industrial-scale reactors. The results reveal that despite temperature-driven dielectric shifts in the column wall and the insulating gap, the normalized sensitivity distribution remains remarkably stable (Pearson correlation coefficient >0.99). This allows room-temperature calibration to be directly applied at high temperatures without compromising reconstruction integrity.
However, raw capacitance measurements are highly susceptible to thermal perturbations within both the packed bed and the column wall. A critical self-dampening phenomenon was identified: granular materials with higher intrinsic permittivity exhibit superior inherent resilience to thermal fluctuations. For example, a uniform 1 % permittivity perturbation applied to baseline values of 2, 6, 10, and 100 yields absolute normalized capacitance deviations of 0.019, 0.009, 0.007, and 0.001, respectively, for adjacent electrode pairs.
Crucially, increasing wall thickness drastically amplifies measurement errors. Numerical simulations showed that for an 8-electrode sensor on a 114 mm diameter column, capacitance-permittivity relationships shift from positive quasi-linear for thin walls to strongly nonlinear and even negative for thick walls. This excessive electric field shielding severely attenuates sensitivity to internal phase distributions.
High-temperature experiments using fluid catalytic cracking (FCC) and silica particles confirmed the numerical predictions. FCC particles (alumina-rich) exhibited stronger temperature-dependent capacitance variations than silica particles, consistent with the dielectric properties of alumina versus silica. The normalized capacitance of the opposite electrode pair decreased from 1.001 to 0.925 as temperature decreased from 600 to 500 °C for FCC, while silica showed much smaller variations.
To improve measurement robustness, excluding adjacent electrode measurements effectively suppresses temperature-induced deviations. For an FCC bed, this extends the allowable temperature fluctuation from about 5 °C to about 16 °C, while for silica, from about 1.2 °C to about 80 °C. This trade-off prioritizes core-region tomographic integrity at the cost of fine-scale boundary resolution.
Finally, the study establishes a theoretical equivalence between thermal distortions and inherent electronic noise, enabling direct application of existing noise-based criteria to define operational thermal limits. These findings provide practical guidelines for designing robust ECT systems for high-temperature industrial reactors.

DOI
10.1007/s11705-026-2680-4