Skip to Main Content
Based on the knowledge representation of cloud theory and rough sets, a rough-cloud model is put forward, which bridges the gap between quantitative knowledge and qualitative knowledge. In relation to classical rough sets, the rough-cloud model can deal with the uncertainty of the attribute and make a soft discretization for continuous ones. A novel approach, including discretization, attribute reduction, value reduction and data complement, is presented. With the origin data rough reduction, a combination cloud generator is put forward, which combines forward cloud generator and backward cloud generator to a close-loop structure. The generator is used to load model for power system to solve the load origin data shortage of distribution system. Considering of the distribution system load data characteristic, the restriction equations and system data complement unit are joined to combination cloud generator, which ensure that the created load data cover most of the system situation without impossible data. The cloud drop reflects the fuzziness and randomness of the load data. The loads are identified by T-S fuzzy model based on the generation cloud drop. The identification result implies the effectiveness and usefulness of the approach by the contrast with some kinds of universal load model.