[en] In pressurized water reactor, ion exchange resins (IER) are used in systems purification. In this thesis, a qualitative study has been performed to predict the behavior of IER while used in nuclear plants conditions. Then, we searched to characterize the IER behavior in column through a quantitative study using analytical solutions. But these solutions worked only for particular cases. In order to find a general solution, we used a new numerical solution: OPTIPUR. To validate this general solution and get a better understanding of the kinetic in column, we performed an experimental study to characterize the resistance to mass transfer in column and to study the sensibility on the parameters influencing this phenomenon. This study is based on the characterization of the initial leakage (initial pollutant concentration at the column outlet). We tested numerous parameters on the initial leakage. We understood the importance of the superficial velocity and indeed of the hydrodynamic conditions on the initial leakage. These numerous results about initial leakage were modeled with an empirical correlation of Dwivedi and Upadhyay in order to validate it. Then, we modeled our results with the two options of OPTIPUR software: option Mass Transfer Coefficient (MTC) and Nernst-Planck (NP). These options encircle experimental results. The MTC option of OPTIPUR gives lower results while those obtained with the NP option are higher than the experimental results. We observed also that only the NP option was valid for a ternary exchange. We proposed solutions to get a better fit with the results obtained with OPTIPUR. We performed other simulations to check the prediction abilities of the software for longer experiments (until the IER saturation). The tendencies were those expected. The OPTIPUR software showed is accuracy and robustness to study column kinetic. (author)