Total And Reactive Load Forecasting

Total And Reactive Load Forecasting

1. Total Load Forecasting

For total load forecast both weather sensitive component and non-weather sensitive component are considered with addition of certain minor component consumption such as public lighting, public water works, sewerage pumping, railway traction, auxiliary consumption in generating station and a reasonable allowance for losses including supply and demand management measurement.

Fig. shows general diagram for total load forecast. The total losses and demand side energy management saving measures are expressed as percentage of the total load.

Sometimes if we know the means and variance of the weather sensitive and non-weather sensitive components then just sum them to obtain the total forecast mean and variance.

On similar basis get the total forecast load density function from weather sensitive and non-weather sensitive components density function.

General Diagram Of Total Load Forecasting

2. Reactive Load Forecasting

So historical data is not good enough to forecast the changes in reactive load as it distribution changes with changes in network configuration and parameters during changing operating conditions.

Due to this reactive loads are difficult to predict as compare to active loads.

Reactive load forecasts are used for security analysis, voltage/ reactive power scheduling etc. To meet the load demand, control of system must manage technical limits of operating components and conditions.

If control is inadequate the structural modifications should be done like placing new lines or new reactive power compensating devices.

Sometimes voltage variation is major problem causing blackouts affecting following factors:

• First swing instability of generators
• Power damping
• Thermal overloads

For good operations up to half of this reactive power is needed to be provided by the generator, synchronous condensers, static VAR compensators or under voltage controlled switched, shunt capacitor banks.

Active load forecast with power factor prediction can give some accurate results, but here accurate historical load data is necessary with assumption of steady state network.