Saturday, October 4, 2014

Forecasting and Planning

Forecasting is to tell what the future would look like given one or more scenarios. Planning is to develop actions in order to reach a desirable stage in the future. Although the forecasting and planning functions often reside in one department of a utility, conceptually mixing the two often results in bad or even fraudulent forecasts and plans. Let me use a set of examples to illustrate the relationship between the two.


The planning department of the Big City Power & Light (a distribution company) need to make a 5 year plan for system maintenance and upgrade. The forecasting team developed a 10-year ahead forecast showing an increasing trend of peak demand at about 2% per year. Majority of the load growth was due to the new tax incentives for attracting high tech companies to a technology park near downtown area. The technology park was expected to add 10,000 high tech jobs over the next 5 years. These jobs would help with the local economy growth.

Example 1: a legitimate case

Based on the forecast, the planning team developed a plan to upgrade the existing infrastructure of the areas near the technology park to serve the load growth due to the newly added jobs.

Example 2: another legitimate case with iterations

After analyzing the power systems near downtown, the planning team concluded that reasonable system upgrade can support up to 1.5% annual growth. It is not economically justifiable to feed the projected demand at 2% annual growth by only upgrading the system. One possible solution was to launch several demand response (DR) and energy efficiency (EE) programs together with the system upgrade. The planning team need to understand what the load growth would look like given different options of the DR and EE programs. Based on the new scenarios provided by the planning team, the forecasting team then developed several new forecasts with different combinations of DR and EE programs. The new forecasts showed annual load growth at 1-1.5% per year. The planning team then developed a plan that includes both system upgrade and selected DR and EE programs.

Example 3: a fraudulent case

One of the high tech companies asked the utility for a new distribution substation to ensure a highly reliable system to serve the data centers. After analyzing the power systems near downtown, the planning team concluded that the annual load growth has to be as high as 3% to justify the addition of the substation. This high tech company lobbied the executives of the utility, who then put a big pressure on the forecasting team. Under the pressure, the forecasting team added two dummy variables to the recent two years with low peak demand to create the 3% annual growth.


An easy way to spot the fraudulent forecast and plan is to check which one comes first. In a legitimate case, the plan should always come after the forecast. Otherwise, it is a fraudulent case. See The Most Valuable Advice I Got from Jim Burke.

Back to Load Forecasting Terminology.

No comments:

Post a Comment

Note that you may link to your LinkedIn profile if you choose Name/URL option.