MTEP16 Chapter 7.3: Independent Load Forecasting
MISO procured an independent vendor, State Utility Forecasting Group (SUFG), to develop three 10-year horizon load forecasts. SUFG provides data used to develop an independent regional load forecast for the MISO Balancing Authority (BA). The first 10-year forecast (2015-2014) was delivered in November 2014. The second 10-year forecast (2016-2025) is due November 2, 2015.
SUFG produces econometric models for 15 states. The SUFG independent load forecast includes a seasonal peak forecast (summer and winter) that is MISO coincident and a coincident forecast for each of the 10 Local Resource Zones. The long-term forecast will be based on MISO Business As Usual (BAU) planning future each year.
The independent load forecast will be a 50/50 forecast, meaning there is a 50 percent probability that the load will either be higher or lower than the forecasted value. The load forecast (demand and energy) for the MISO BA will be forecasted for each state, and then aggregated into each MISO Local Resource Zone (LRZ) through the use of allocation factors. The MISO BA has 36 Local Balancing Authorities (LBA). The LBAs are aggregated into ten Local Resource Zones (LRZs) (Figure 7.3-1).
The independent load forecast is not intended to replicate or replace an individual Load Serving Entity (LSE) or Transmission Owner (TO) forecast. This is an independent and transparent approach to develop a MISO load forecast that relies on publically available data, limiting dependence on confidential or vendor data and new data requests. Each state forecast model and the associated assumptions will be made available to stakeholders, and will require no vendor-specific software. SUFG is using common industry econometric forecast data and software (Global Insight, EViews).
Project Schedule and Deliverables
This project is a three-year effort (Figure 7.3-2), with forecast deliverables due annually at the beginning of November 2014, 2015 and 2016. Key activities and milestones are outlined for the 2016-2025 forecast (Table 7.3-1).
The scope of the 2016-2025 forecast was updated based on stakeholder feedback received in the first quarter of 2015. LRZ 10, previously a part of LRZ 9, was added in Mississippi. SUFG updated state econometric models and the conversion of the energy forecast to the peak forecast. SUFG also modeled multiple weather stations in the state econometric models, as well as improved modeling of demand response, energy efficiency and distributed generation. Finally, SUFG incorporated uncertainty in the drivers of the econometric models into the high and low forecast bands by estimating confidence intervals based on the historical variance of the drivers.
MISO also made progress on a load forecast comparison between the Independent Load Forecast and the Aggregated LSEs Forecast. The objective of this comparison is to identify where the forecasts differ in order to determine if model, methodology or inputs can explain these differences. The load forecast comparison does not test whether one forecast is more accurate than the other; the goal is to understand where and why there are differences. Data inputs that explained some of the differences were identified. MISO used historical energy and demand data from 2010 to 2014 to attempt to put forecast starting points and trends in perspective. Since forecasts assume normal weather, this MISO historical data was then weather normalized so that historical data without the effects of weather would be available..
|Key Activities And Milestones||Target Dates|
|2016-2025 Independent Load Forecast||11/1/2015|
|Stakeholder Workshop #1 – Review 2015 project plan, discuss potential improvements, load forecast comparison||1/22/15|
|Stakeholder Comments Due||2/3/15|
|Acquire (update) state level historical data||3/2015|
|Update econometric forecasting models for each state||4/2015|
|Stakeholder Workshop #2||4/23/2015|
|Stakeholder Workshop #2 Comments Due||5/14/2015|
|Determine allocation factors to convert state energy forecasts to each Local Resource Zone forecast||6/2015|
|Review energy to peak demand conversion model for each Local Resource Zone||7/2015|
|Incorporate econometric model drivers||6/2015|
|Generate a 10 year annual energy forecast for each state using its econometric forecast model||7/2015|
|Stakeholder Workshop #3||7/23/2015|
|Stakeholder Workshop #3 Comments Due||8/6/2015|
|Determine 10 year annual energy forecast for each Local Resource Zone||8/2015|
|Determine 10 year seasonal peak demand for each Local Resource Zone||8/2015|
|Determine MISO’s 10 year forecast for annual energy and seasonal peak demand||9/2015|
|Stakeholder Workshop #4 – Review 2016-2025 Forecast results||9/17/15|
|Stakeholder Comments Workshop #4 Due||10/8/15|
|Independent 10 year (2015-2024) Demand and Energy forecast report completed||11/2/15|
|Stakeholder Comments Due||11/13/15|
Table 7.3-1: Independent Load Forecasting Project detailed project schedule 2015.
The MISO transmission system needs to be planned such that it is prepared for changes in the resource mix caused by changing environmental regulations, commodity prices, renewable integration and economic conditions.
More than 141 LSEs and approximately 41 TOs submit demand forecasts annually; each with potentially different assumptions and methodologies. Each LSE and TO uses its own parameters, making it impossible to develop a MISO region-wide load forecast based on a common set of economic conditions for scenario analysis in long-term studies. An unaccounted-for deviation in a load forecast can result in increased reliability risk from the industry reliability standard (one day in 10 years) because it is difficult — if not impossible — to understand the drivers and changes in an aggregated bottom-up, long-term forecast.
A single, MISO region-wide load forecast can be viewed as a top-down approach for the region; it has the benefits of one set of assumptions, and can be used in other regional studies and future analysis. This top-down approach for load forecast fits in with MISO’s “Top Down, Bottom Up” transmission planning process.
This is an alternative forecast methodology. It is not intended to replicate or replace each LSE’s or TO’s forecast process. MISO will continue to use the load forecasts provided by the LSEs and TOs in MTEP and Module E: Resource Adequacy as required by the MISO Tariff.