Power equipment can be maintained more cost effectively by planning based upon their condition. To estimate the condition of a plant, a number of sensors can be used to measure data such as temperature and pressure, which are then collected by means of an information system along with processes and storage. For this purpose, maintenance staff must install software into the information system each time a sensor is installed; the greater the number of such tasks, the less efficient the maintenance. This paper proposes a method to facilitate the installation of software and thus reduce the burden on maintenance staff. The method proposed manages information with respect to software classes, creates instances from the classes according to sensor detection, and activates them based upon related instance states. Results using an experimental system show that the method works as we expected and also seems to be scalable.