By David Kligman
SAN FRANCISCO — Electricity is a must have, as nearly any Fresno customer would tell you when an air conditioner provides relief from sweltering summer heat.
For a utility like PG&E, providing safe, reliable and affordable power is critical. But doing so for millions of Californians requires being able to forecast where and when that power will be needed, and its impacts on the electric grid. To be accurate, that forecast also must include the positive contributions made by customers who conserve energy and renewable forms of energy.
Every fall, PG&E spends up to two months analyzing how much electricity will be needed over the next several years throughout the company’s 70,000 square miles.
This year, PG&E is using new software that’s helping the utility forecast its power needs. That tool, which PG&E helped customize with a Cincinnati software developer, uses customers’ SmartMeter data to make the forecasts even more precise.
The calculations are highly specialized. And they have enormous implications on how PG&E spends its customer dollars and on which projects.
Better long-term forecasting means PG&E’s millions of customers will benefit from improved operational reliability and planning, and an increased ability to incorporate more renewables and cost effective resources.
Forecasting means improved operational reliability
Ultimately, improved long-term capacity planning leads to lower rates and improved reliability, said Manho Yeung, PG&E’s senior director for system planning and reliability.
“Use of this technology demonstrates how PG&E is using the latest tools to ensure our capacity-related capital investments minimize costs to our customers,” Yeung said.
The search for a new forecasting tool began several years ago. PG&E met with various software companies and ultimately began a contract with Integral Analytics’ LoadSEER platform because of its ability to integrate SmartMeter data.
Prior to the new software, PG&E used Excel spreadsheets that were updated each year by local engineers. The results were more of a simulated forecast based on historical electric load data over several years.
The new analysis looks at history but also takes into account PG&E’s planned construction projects, zoning permits, distributed generation and SmartMeter data. And it’s now in a secure database that’s easier for PG&E planners to access.
“The biggest advantage is that the program forecasts down to the circuit level and the bank level,” said Donovan Currey, a Chico-based senior area planning engineer. “Our old forecast methodology used a geographic area and the engineer had to use his judgment to place the load where he thought it might happen. This program is putting it where it’s expected to happen.”
Accurate forecast of July heat wave electric load
Currey pointed to the July heat wave in California. The software forecasted PG&E’s electric load on many circuits within 1 to 2 percent of actual loads, much more accurately than the utility’s previous forecasting program.
LoadSEER is helping PG&E determine exactly how many transformers and power lines will need to be upgraded. It looks at variables including how the increased use of electric vehicles will impact the power grid.
“Accurate forecasting is fundamental to identifying where and when capacity infrastructure improvements are required to provide safe and reliable service to PG&E customers,” Yeung said.
PG&E hopes to make further refinements to the software and is working with the Ohio company on additional features, including an important agricultural load forecasting component. This would take into account factors unique to farmers — commodity pricing and the availability of water — that will improve overall forecasting.
Beginning this month, about 80 PG&E engineers from Eureka to Bakersfield will spend about six to eight weeks using the software. It will include all the data from the past year, including this summer’s peak loads. PG&E will finalize the results on Dec. 15 and use the information to plan for 2014 and beyond.
Email David Kligman at David.Kligman@pge.com.