Agile chemical plants operate in a number of states including steady states and frequently switch between them. Different control configurations or controller parameters may be used for control of the process in different states. The obvious need for efficient and automatic identification of the different process states using large historical data sets, in lieu of manual annotation by an engineer, provides the motivation for this work. Traditional clustering methods are computationally expensive and normally perform poorly on temporal signals. A two-step clustering method based on principal component analysis (PCA) is proposed in this paper. Process states are first classified into modes corresponding to quasi steady states and transitions. A novel multivariate algorithm is used to segment historical data into modes and transitions. Dynamic PCA-based similarity measures are then used in the second phase to compare the different modes and the different transitions and cluster them. This two-step methodology can be applied directly to multivariate process data and has low computational requirements. Extensive testing on a fluidized catalytic cracking unit and the Tennessee Eastman process simulations illustrate the effectiveness of the proposed method.