Profiling consumers in a water distribution network using K-Means clustering and multiple pre-processing methods.