Medium range forecasting of hourly power system load by time series analysis using the Walsh transform

摘要: Abstract Medium range forecasts of hourly load demand for a span 24 h to 168 (one week) are required the preparation short term maintenance schedules unit auxiliaries and peaking stations apart from maintaining security constraints minimizing operational costs. Forecasts daily seven days by available multiplicative SARIMA model suffer divergent error levels multistep ahead forecasts, take all features any particular day, order such models will be prohibitively large. In present paper unique approach has been made time series analysis stochastic processes: multigrade periodicities in curve have gainfully exploited grouping data subgroups characterizing each week day separately then group is replaced an orthogonal transform known as Walsh transform. The amplitudes transformed spectrum modelled Box-Jenkins taking inverse (Walsh) transforms.