The ex situ decontamination of uranium polluted soils was performed by alkaline washing using mechanical agitation and ultrasound field. Two types of polluted soils were considered in terms of texture and organic content. The tests were performed using experimental design: a 23 factorial plan for mechanical washing and Taguchi L18 orthogonal matrix for ultrasound (US)-assisted decontamination. The factors considered in mechanical washing were temperature, duration, and pH. For ultrasound decontamination, the US frequency was added. The optimum was estimated based on statistical analysis and validated by confirmation experiments. The study revealed that in ultrasound field, the decontamination degree is increased with over 25% compared with mechanical stirring, while the duration of the process is drastically reduced (from 120 to 25 min). The most important factor leading to the increase of decontamination is the ultrasound frequency. To refine the result, artificial neural network modeling and subsequent optimization by genetic algorithms were also carried out for the decontamination in ultrasound field, and new optimum operating conditions were identified and validated. The best operating conditions identified were temperature around 50 °C, pH = 8, and ultrasound frequency of 24 kHz. In these conditions, the decontamination degree reached 85% for Soil 1 (characterized by low clay and organic content) and 69% for Soil 2 (with high clay and organic content).
Keywords: Artificial neural network; Experimental design; Optimization; Soil washing; Ultrasound field; Uranium decontamination.