Main Capabilities

Unlike other options which involve manual processes and data links (spreadsheet approach), aggregated variance levels and settlements (top-down approach) or large up-front investments (packaged systems approach), ELM uses a bottom-up approach to deliver the highest levels of customization, integration, data cleaning and reporting.

ELM handles all the key areas involved with Load Forecasting customer load of deregulated entities:

  • Customer capture (EDI, market, and client)
  • Usage cleaning, profiling, and weather normalization
  • Custom profiling/modeling (neural network, regression, similar-day picker)
  • Load forecasting (10-Day short term / multi-year long term)
  • Load settlements capture and comparison to forecast/backcast
  • Data aggregation and reporting support

Data Capture

ELM captures and supports the key data required for Load Forecasting:

  • Captures enrollment, drop and usage data directly via EDI transactions 
  • Product codes provided by client at the meter level with start/end date 
  • Maintains a history of customer attributes such as meter type change
  • Ability to forecast losses separately as either fixed or index
  • Supports complex products
  • Captures standard/custom block products   

Short-Term Load Forecast

ELM produces short-term forecasts using forecasted weather on a daily basis, delivered directly to the Wholesale Supply User Team:

  • One- to 10-day hourly forecast by LFG and market channel segments, includes multiple loss components
  • Uses short-term weather forecast from preferred vendor
  • Option to include weather bands to assess load swing risk (+/- five (5) degrees)
  • Issued daily, typically by 7 a.m. central time
  • Reports available at a customer-group level or rolled up to zonal level or total company
  • Used for scheduling load and revising supply positions; can modify to support revenue/margin forecasting

Long-term Load Forecast  

ELM produces multi-year long-term forecasts with normal weather:

  • Weekly forecast based on normal weather view
  • Based on committed load – current customers and contract end dates 
  • Can incorporate client-defined views of customer growth and attrition
  • Reports available at customer group level or rolled-up to zonal level or total company
  • User is able to define length of forecast and normal weather definition
  • Used for scheduling load and revising supply positions; can modify to support revenue/margin forecasts    

Backcasting

ELM produces backcast load using actual weather:

Provides refined views of projected actual load and drives variance calculations:

  • "After-the-fact" forecast using actual weather and same tools 
  • Initial – change forecasted with actual weather 
  • Adjusted – change weather, counts, usage factors   

Forecast accuracy and variance measurement:

  • Calculates hourly forecast variance by segment 
  • Import client’s settlement data (ERCOT/PJM) 
  • Disaggregates variance drivers – weather, count, usage 
  • Provides ability to assess risk and imbalance costs  

Platform Integration

ELM integrates with other platforms:

  • Currently integrated with ECI and ESG for EDI data capture; can integrate with any EDI vendor
  • Custom modeling supported by a separate external application (MATLAB)
  • Forecasted and actual weather data from WSI is included in the service cost
  • Customer implementation can integrate with any weather vendor