CMPAS conducts load forecasting to help utilities plan their future.
Short-term Load Forecasting
CMPAS uses a neural network forecasting tool called Nostradamus. CMPAS uses the results of this software to forecast the hourly load to schedule in the MISO Energy market each day. In compliance with MISO requirements, the Agency submits a seven-day hourly load forecast.
The forecasting tool works by having the user input large quantities of hourly historical data such as peak electrical demand, temperature, humidity, wind speed, and day of the week. The model then identifies correlations between historical inputs and uses them in conjunction with weather service data to forecast future hourly loads.
Long-term Load Forecasting
Using weather normalization statistical techniques, CMPAS produces long-term forecasts that range from 1-20 years.
For a one-year forecast, CMPAS uses a more conservative 90/10 confidence level forecast to ensure members have adequate capacity resources to meet their coming year MISO regulatory reserve margin obligation. For intermediate and long-range resource planning, a 20-year 50/50 confidence level forecast is used.
Both of these long-range forecasts are created using an economic forecasting model that requires the user to input census and demographic data.
Natural Gas and MISO Market Price Forecasting
In addition to performing load forecasting, CMPAS also hires an independent consultant to perform a natural gas price forecast and a MISO market forecast for future capacity and energy prices.
Comprehensive, long-range planning efforts depend on these forecasts, and CMPAS uses them to evaluate the request for proposal (RFP) submissions for “build versus buy” decisions.