In the field of power equipment condition evaluation

the concept of equipment fault is obviously fuzzy

so the evaluation results are difficult to meet the requirements of differentiation

refinement and accuracy required by digital and intelligent power systems. In this paper

the concepts of fault and fault probability are clarified and expanded

the difficulty of sample collection is reduced

and the influence of maintenance strategy on the statistical results of fault probability is overcome. Based on the four-element stress- strength interference model

a fault probability calculation method including the defect inducements and their influences is constructed to solve the problem of the characterization of multi-dimensional influences with complex stochastic course. Based on the correlation between condition indicators and defect

a fault probability calculation method grounded in condition indicators is derived. Taking the calculation of overheat fault probability as an example

the implementation process of the algorithm are shown. Case calculation results show that the algorithm is feasible and stable

and it is an effective way to achieve differentiated and refined evaluation of equipment condition.