Skip to main content

Overview

This section will discuss how to run a short-term forecast using Apollo's Command Sequence.

  • Required Inputs
  • Step-by-step procedure
  • Exercise

Short-term Forecasting​

Forecasting is essential to electricity market analysis, providing analysts with insights into potential market conditions based on various assumptions.

Prerequisites​

To perform a Short-Term Forecast in Apollo, the following inputs are required:

  • Demand File: Forecasted load requirements.
  • Must-Run Generation Capacity Factor File: Forecasted Must-Run Generation data.
  • Bid Files: Forecasted bid profiles (Historical, Typical, or Custom).
Modeling Outages

To improve model accuracy, consider including a WAPOS event file for scheduled outages or enabling Monte Carlo Simulation for random outage modeling.

Tutorial Video (to add)​

1

Forecasted Demand​

iLoad Demand Forecasting Service

The iEnergy iLoad Forecasting Service streamlines this workflow by providing rolling Demand and MRG Capacity Factor files pre-formatted for direct use in iPool and Apollo. This service is powered by a sophisticated hybrid intelligence framework that integrates premium, high-resolution weather data with advanced machine learning and AI to provide high-fidelity energy predictions for the Philippine power grid.

  • Advanced Demand Forecasting: The system processes a 70-feature matrix that includes cyclical temporal encodings, Philippine holiday detection, and "human comfort" drivers like the Heat Index and Cooling Degree Hours (CDH).
  • Continuous Calibration: To ensure sustained accuracy, the service features Rolling 14-Day Auto-Calibration.
Manual Demand Forecasting

Method 2: Manual Forecasting (No Subscription)​

If you do not have an active iLoad Forecasting subscription, you must manually generate your forecast files using one of the following options:

  • Option 1 (Historical Persistence): Utilize historical Demand data by applying previous Demand.csv values to your target dates.
  • Option 2 (Manual Correlation): Manually approximate demand by correlating historical output with daily max/min temperature forecasts.
2

Forecasted Generation​

iLoad Generation Forecasting Service
  • Wind Capacity Factors: For wind stations, the utility utilizes premium 100m hub-height wind data.
  • Solar Capacity Factors: Solar forecasting is driven by a Physics+ML Hybrid model.
  • Continuous Calibration: Features Rolling 14-Day Auto-Calibration.
Manual Generation Forecasting

Method 2: Manual Forecasting (No Subscription)​

If you do not have an active iLoad Forecasting subscription, you must manually generate your forecast files using one of the following options:

  • Option 1 (Historical Persistence): Utilize historical Must-Run Generation data by applying previous MRHCFac.csv values.
  • Option 2 (Manual Correlation): Manually approximate solar generation by correlating historical output with daily max/min temperature forecasts.
Note

If multiple scenarios are present, the top-most scenario in the list will be executed first. When new scenarios are added to Apollo, you must click the Restart button to ensure proper initialization.

Exercise​

Give it a try and do this exercise:

Step 1 of 4

Backcast Calibration

Execute a one-week Historical Load simulation
Perform Backcast Calibration to establish a baseline