Considering the problems of the uncertainty of the load type in household scenario and the incompleteness of the load signature database in the non-intrusive load database, which can easily lead to the decrease of the accuracy in load identification, this paper proposes a load identification method to cope with these problems. On the base of electrical signatures, this method also use time signature which includes the characteristics of the length of operation time, load operation time, working period and vacation. In this method, firstly, we use the piecewise- normalization mean-shift clustering method to cluster the detected load event features and obtain the number of potential load types. Then we count the time signature and power signature of load events to get their probability. And the Bayesian method is used to identify the load by decision-making. Finally, this paper uses the AMPds public data set to do the actual test, the experimental results show that this method has the good identification effect to this scene.
