Industrial thermal processes generate a large amount of gases to be treated before released to atmosphere. Filtration is one of the major steps of this treatment as it ensures no release of solid particles. In any given field, electrostatic precipitators are technologies of choice because of several reasons they avoid significant pressure drop in industrial systems; they do not need any filtering media; they may have simple geometry; their cleaning is easy; These reasons explain why this technology is particularly attractive when the thermal treatment is applied to radioactive material and then why the CEA works for several years to improve electrostatic precipitation applied to its needs. The main objective of the research and development performed in this field was to increase the filtering efficiency of the filters together with their endurance. After more than ten years of work, objective has been broadly reached and can be applied to various applications, nuclear or not. The major purpose of the present paper is to explain the different steps of research that have led to pass from a few tens of minutes of optimal efficiency up to several hours. Electrode geometry, electrostatic field modeling, high-voltage direct and pulsed current, sonic and ultrasonic cleaning were at the heart of these developments and have led to very conclusive results.