2018

10-Year
Assessment

Load forecasting methodology

ATC will initially use load forecasts provided by the company’s end-use load-serving customers. In general, customers are required, to provide ATC with monthly peak demand forecasts for the next 11 years. ATC may, in the future, develop load forecasts either concurrent with or independent of the company’s load-serving customers. In addition, ATC may, in coordination with the company’s load-serving customers, develop representative load duration curves based on actual and normalized load conditions.

In utilizing or developing load forecasts, the following methodology will be used:

  1. Summer peak demand forecasts will be calculated in such a way that there is an almost equal probability of exceeding or falling short of the forecast when average peak making weather does occur.
  2. Shoulder maintenance window: In order to develop a shoulder maintenance window model, a maintenance window analysis should be performed for the ATC footprint periodically. This analysis should determine:
    1. How many load pockets the ATC system should be divided into.
    2. What the overall load level in terms of a percentage of the summer peak should be achieved for each load pocket.

    Then for each load pocket, the shoulder maintenance window forecasts will be developed such that the scalable loads are scaled to a pre-calculated percent of the summer peak demand forecasts while holding the non-scalable loads smaller than or equal to 5 MW constant and applying shoulder load ratios[1] for the non-scalable loads greater than 5 MW. The resultant overall load level should meet the target determined in the latest maintenance window analysis. The ratio of the real to reactive power of the loads will remain unchanged from the summer peak ratio.

    [1] To enhance the modeling of shoulder and light load conditions for the ATC Planning analysis, during the load forecast process, ATC requested local distribution companies to provide shoulder-to-peak ratios and light-to-peak ratios for the non-scalable loads greater than 5 MW.

  3. Winter peak demand forecasts will be developed such that the scalable summer peak loads are scaled to Local Distribution Company (LDC) chosen percentages for the following December.  Non-scalable loads remain unchanged. The ratio of the real to reactive power of the loads will remain unchanged from the summer peak ratio.
  4. Fall/spring off-peak demand forecasts will be developed such that the scalable loads are scaled to LDC chosen percentages for the month of November for fall or the month of April for spring. Non-scalable loads smaller than or equal to 5 MW remain unchanged while applying shoulder load ratios for the non-scalable loads greater than 5 MW. The ratio of the real to reactive power of the loads will remain unchanged from the summer peak ratio.
  5. Summer 90/10 peak demand forecasts will be developed that reflect above-average summer weather and peak demand conditions. A true summer 90/10 forecast at the ATC aggregate load level will be developed in such a way that there is a 90 percent probability of falling short of and a 10 percent probability of exceeding the forecast due to weather conditions. Summer 90/10 peak demand forecast will then be developed such that the scalable loads are scaled to a pre-calculated percent of the summer peak demand forecasts while leaving the non-scalable loads unchanged. The resultant overall ATC load level should meet the determined 90/10 forecast. The ratio of the real to reactive power of the loads will remain unchanged from the summer peak ratio.
  6. Light load (50 percent of summer peak) demand forecasts will be developed such that the conforming loads are scaled to a pre-calculated percent of the summer peak demand forecasts while holding the non-scalable loads smaller than or equal to 5 MW constant and applying light load ratios for the non-scalable loads greater than 5 MW. The resultant overall ATC load in the Light load model is  approximately 50 percent of the summer peak. The ratio of the real to reactive power of the loads will remain unchanged from the summer peak ratio.
  7. Minimum load (40 percent of summer peak) demand forecasts will be developed in two steps:
    1. The scalable loads are scaled to a pre-calculated percent of the summer peak demand forecasts, while holding constant the non-scalable loads smaller than or equal to 5 MW constant and applying light load ratios3 for the non-scalable loads greater than 5 MW. The resultant overall ATC load in the Minimum load model is approximately 40 percent of the summer peak.
    2. Historical EMS model data are used to help developing the reactive load forecast in the minimum load models.
      • Obtain the historical (Easter Sunday 4:00 AM and Memorial Day 6:00 AM) ATC control area reactive load data for at least three years from EMS models and average them to get the reactive load target for each control area.
      • For each of the ATC control areas, the scalable reactive loads are scaled to a pre-calculated level of the summer peak demand forecasts, while holding the non-scalable loads smaller than or equal to 5 MW constant and applying light load ratios3 for the non-scalable loads greater than 5 MW. The resultant overall reactive loads for each control area should meet the target determined in the step above.