A Fast Regression Testing Method Using for EMS System Engineering Project

Abstract

EMS(Energy Management System) is a software system with complex structure, which is used in power grid dispatching automation. In the life cycle of the system, it will go through a number of tests, such as FAT(factory test), SAT(acceptance test), etc, whose test cases will be as many as ten thousand. In the maintenance of EMS system, requirements may be often changed, which motivates the component modification to create new versions and their accompany tests. Regression testing strategies aim at the retest for modification affection in EMS software maintenance. This paper presents a EMS system regression testing method based on a new Component Testing Association Model. In this method, previously executed test cases are selected to generate the minimal regression test suite by the identification and impact analysis for the modification-affected component groups. Compared with traditional methods, our approach focus on the complicated interaction between components, which is more applicable to component system regression testing. Our method greatly saves the EMS system regression testing costs which go through the functional adjustment. In practical projects the execution of a moderate scale regression test was reduced by about 2 weeks.