A beam-induced damage to the Tevatron collimators happened in December 2003 was induced by a failure in the CDF Roman Pot detector positioning during the collider run. Possible
scenarios of this failure resulted in an excessive halo generation and superconducting magnet quench have been studied via realistic simulations using the STRUCT and MARS14 codes. It is shown that the interaction of a misbehaved proton beam with the collimators result in a rapid local heating and a possible damage. A detailed consideration is given to the ablation process for the collimator material taking place in high vacuum. It is shown that ablation of tungsten (primary collimator) and stainless steel (secondary collimator) jaws results in creation of a groove in the jaw surface as was observed after the December's accident. The actions undertaken to avoid such an accident in future are described in detail.