 Risk
Analysis
Most T&D planners have historically used SCADA data or other metered
data to judge the extent to which their circuits or feeders peak during
the course of the year. Sometimes the reliance on this type of historical
data is appropriate (e.g., no growth on the feeder, non-weather sensitive
load). Other times, these historical measurements are taken during mild
weather, or do not provide a complete picture of the risks of reaching
higher peak loads during times of more extreme weather.
This is the forte of Load At Risk Analysis (LARA), which provides
a complete set of predictions of feeder and/or spot loads under all
known historic weather conditions going back 10 to 50 years. Not only
is it possible to reasonably estimate the circuit’s load for the coming
year (weather normal peaks), but it is also possible to derive a full
distribution of expected loads under ALL weather conditions. With this
information, planners can set their risk tolerances (or planning temperatures)
at different levels for different feeders, depending on where each feeder
reaches its own peak conditions.
The way that LARA achieves this level of specificity is through its
ability to construct optimum non-linear predictions of load as a function
of several input variables (e.g., wind speed, precipitation, humidity,
temperature). The best non-linear response function is selected, among
hundreds for each hour, for each month, for each weekday and weekend.
Using these optimally selected weather response functions, millions
of simulations are calculated for each hour for each feeder and/or spot
load to arrive at a full set of predictions of loads and the load at
risk of not being served.
LARA simulations and analysis are applied to KW, KVAR, and KVA information.
Output results are summarized and graphed in color to provide a quick
and informative view of each feeder under study. Spot loads can be defined
by the users, and loads as small as 15 KW are just as easy and quick
to model as large loads. Similarly, class loads can be aggregated and
analyzed as groups, where the expected weather responses are thought
to be similar in nature.
Highlights
- Estimates full distribution of loads over all weather conditions.
- Hundreds of weather responds function tested at each hour, with
the best performing function selected.
- Virtually any size load can be modeled, with spot loads and feeders
being the most common.
- Output provided in easy-to-read color graphs and summary tables.
|
