Computational model‐based design of molten salt electrorefining process for high‐purity zirconium metal recovery from spent nuclear fuel

Summary Effective decontamination methods for spent nuclear fuel (SNF) cladding are required to recycle Zr, which is a valuable resource, by separating high purity Zr from the actinides. In this study, computational modeling is performed on an electrorefiner to achieve high‐purity Zr metal recovery without ZrCl and U from SNF Zircaloy‐4 cladding utilizing a commercial fluid dynamics code. A three‐dimensional (3D) electrorefining model specialized in simulation of multistep electrochemical reduction (eg, two‐step reduction of Zr(IV) to Zr via ZrCl) is developed by coupling the numerical models of the Butler‐Volmer equation and fluid dynamics. This model is validated by benchmarking the chemical formula of cathode deposits obtained from lab‐scale electrorefining experiments utilizing fresh Zircaloy‐4. The computational results are consistent with the compositions of Zr and ZrCl in the cathode deposits, depending on the initial ZrCl4 concentrations. Based on the developed 3D model, a pilot‐scale electrorefiner for the SNF cladding is simulated with several derived design parameters. The effects of rotating anode and cathode, potential range, molten salt weight, and the number of anode baskets are determined to optimize the electrorefiner design to achieve the suppression of ZrCl and U codeposition. The electrorefiner throughput when employing the optimized design and operating conditions is predicted to be 0.1 to 0.2 ton/y, as only pure Zr metal is recovered. 3D computational modeling is performed on electrorefining to achieve high‐purity Zr metal recovery without ZrCl, U, and alloying elements from spent nuclear fuel Zircaloy‐4 cladding. The electrorefining model for the simulation of Zr multi‐step reduction is developed by utilizing the Butler‐Volmer equation with fluid dynamics and validated by benchmarking Zircaloy‐4 electrorefining experiments. Pilot‐scale electrorefiner is simulated with optimization of design parameters, and throughput is shown to be 0.1 to 0.2 ton/year for pure Zr recovery.