Forecasting Service

Forecasting customer usage, whether for scheduling, position management or market assessment, is one of the most difficult but critical tasks for a retailer. REBO has extensive experience developing custom forecasting systems that take advantage of REBO’s data cleaning and management system. We offer everything from our own forecasting system and custom enhancements that meet your specific needs to custom system development and deployment.

REBO System

REBO’s forecasting system provides four forecasts:

• Short-term scheduling forecasts (up to 10 days) based on daily and hourly weather forecasts. Weather data is updated daily.
• Intermediate term forecasts (10 days to 3 months), based on 30-year average weather with an adjustment factor based on National Weather Service (NWS) monthly and seasonal temperature forecasts. NWS forecasts are updated monthly approximately mid-month.
• Long-term forecasting (greater than 3 months), based on 30-year average weather. The 30-year average weather data is updated annually.
• Dictated nominations for gas.
• Forecasts can be based on utility or ISO provided equations or custom forecasting equations generated by REBO.

REBO’s forecasting system produces interval forecasts for any period from one day to many years. Intervals are defined by the utility or ISO in which customers are being served. Short-term forecasts are reported by interval and long term forecasts are reported by interval and as calendar month summaries. You can also specify the form and format for additional forecast reports.

REBO offers automated forecasting for both electric and natural gas retailers, either as an ASP or on your servers.

In operation, each short-term forecast is run overnight as soon as utility or ISO daily data is posted and is available by 7 am to your forecasting, supply management or scheduling group. They can access the forecasts, accept or reject the forecasts, and either rerun the forecasts or schedule from them. The primary reason for rejecting and rerunning a forecast is a significant change in the weather forecast.

Longer-term forecast are run ad hoc by the client, generally for long term position management when there is a change in its forecasted customer base or when entering new service territories. Forecasting for a new service territories also requires set-up of the utilities and ISO in the new territory.

Each non-interval customer or customer group will have a set of temperature-sensitive equations. The equations may be those provided by the utility or ISO or custom equation sets developed through REBO’s proprietary process. Each equation set is solved using forecasted weather data and then scaled by each customer’s prior year usage.

For utilities or ISO’s that do not provide forecasting equations, REBO offers custom equation development. The REBO equation sets are made up of 288 or more equations defined at the season, day type and day or hour level.

For interval customers, two options are available. Since manufacturing customer usage is generally not weather-sensitive, the forecasting system looks at prior day, day-7, or same day prior-year and determines the best fit for forecasting. Non-manufacturing customers are weather-sensitive, so either the set of custom weather-sensitive equations can be developed for forecasting or the same process used for manufacturing customers can be used.

Utility and ISO transmission and distribution losses are embedded in aggregate forecasts.