To determine the alarm threshold for criticality accident alarm systems (CAASs) used in facility monitoring, it is necessary to select appropriate benchmark experiments for validation. This paper constructs a simplified model and a representative virtual equipment room model, taking into consideration the characteristics of CAASs in the equipment room. The functionality of calculating sensitivity coefficients using the differential operator sampling approach in the Monte Carlo code RMC was implemented to solve the fixed source problem. Following that, a sensitivity and uncertainty analysis was conducted to examine the similarities between systems and explore the factors influencing them. The results indicate that the neutron source term spectrum and the concrete material significantly affect the similarity between systems. Additionally, the thickness of the concrete between the detector and the inner wall has a moderate influence on similarity, whereas the geometric position of the source term has a negligible impact. The impact of the spent fuel composition on similarity can be neglected. Based on the similarity analysis procedure proposed in this study, it is concluded that the simplified model exhibits a relatively high similarity to the virtual equipment room model. Hence, the research conclusions based on the simplified model are applicable to the virtual equipment room model, offering valuable insights for similarity studies of CAASs in engineering.