PowerApps Short Term Load Forecast
The objective of short term load forecast is primarily for purpose of operational planning and control.
The load forecast range could be in the range of few hours to few days.
Forecast is needed for the total load on the system as well as area wise load estimations and in many cases the bus wise load estimation for important buses under consideration.
The output of the short term load forecast is used for
- Unit commitment
- Planning for peaking units
- Economic Dispatch
- Checking feasibility of operation using load flow studies
- Security assessment for planned system operation
Many factors influence the system operating loads as follows
- Time of the day
- Type of day, whether special day, holiday, rainy day, week day, weekend etc
- Type of season, summer, winter, rainy etc
- Expected temperature, humidity, rainfall, snowfall, etc
The Power Foundation Classes [PFC] of PowerApps has standard built in libraries that handles simple trend type to regression models from historical data of the hourly loads and other regression model parameters. The load estimation at bus level may also be considered as influenced by bus voltage and bus frequency as additional regression parameters.
An important aspect of PowerApps load forecast models is the modeling of the “noise” , which is the difference between the historical information and the actual model. With this noise model the original forecast model is refined further providing better mathematical model. However, it should be noted , irrespective of accuracy of load forecast model used , there is no guarantee in any load forecast model, that the forecasted will predict accurately the future loads without any errors. Consequently some degree of margin needs to be provided in the predicted system and bus wise loads for purpose of planning exercise.