Electricity customer classification using frequency–domain load pattern data

摘要: Abstract In competitive electricity markets, customer classification is becoming increasingly important, due to new degrees of freedom the providers have been given in formulating dedicated tariff options for different classes. Several techniques proposed literature, which load patterns are typically represented by time–domain data. However, a good pattern representation requires using several data each customer, causing possible difficulties storing large amount company's databases. order reduce number be stored an original solution this paper, based on post-processing results measurements obtain reduced set defined frequency domain. The successively used procedure, e.g. suitable clustering technique, whose adequacy can assessed means properly indicators. This paper provides mathematical background frequency–domain definition and investigates effectiveness choices stored. Results obtained customers belonging real distribution system presented discussed. These show that effective reducing while maintaining satisfactory level adequacy.