Skip to content
  • Ron Flax of boulder is checking his smart meter of...

    Ron Flax of boulder is checking his smart meter of the house. Boulder, CO. Friday. October 5, 2012. The costs Xcel Energy's SmartGridCity pilot in Boulder spiraled even as the utility failed to deliver on some of the promises of the project. Hyoung Chang, The Denver Post

of

Expand
Denver Post reporter Mark Jaffe on Tuesday, September 27,  2011. Cyrus McCrimmon, The Denver Post
PUBLISHED: | UPDATED:

The idea of a “smart grid” — a computerized system to control energy from power plant to the kitchen dishwasher — was sweeping the utility industry in 2007, and Xcel Energy planned to launch the world’s biggest project.

The company’s SmartGridCity would manage power flows, allow more wind and solar on the grid, and enable consumers to control electricity consumption. Xcel chose Boulder for the ambitious plan.

Five years later, few of the promises are fulfilled. Costs nearly tripled to $44.5 million, and Xcel wants its Colorado customers to foot the bill.

Xcel executives say the project provided valuable lessons in managing power.

“We think it has had a lot of successes — granted, they are successes most of our customers do not see,” said Karen Hyde, vice president for rates and regulatory affairs at Xcel subsidiary Public Service Co. of Colorado.

Critics call SmartGridCity poorly planned, poorly managed and a failed experiment.

“You didn’t have to spend $44 million to learn what Xcel did,” said Tim Schoechle, a Boulder-based smart-grid analyst.

The Denver Post reviewed thousands of pages of company documents and transcripts of testimony filed with the Colorado Public Utilities Commission and interviewed Xcel executives to piece together a picture of what happened with SmartGridCity.

The Post found that Xcel:

• Struggled with rising costs from the outset — the company’s finance committee learned in May 2008 that the utility’s share of the tab could double — but did not inform regulators about the overruns for a year.

• Installed a $21 million broadband fiber-optic network linking homes, substations and central control in Boulder, an expensive technology no other U.S. utility has adopted and that the company says it won’t use again.

• Lost support from partner companies, which were supposed to provide technical assistance and financial backing.

• Had trouble developing in-home energy devices and ended up installing just 101 in Boulder instead of the targeted 1,850.

• Abandoned the idea of companywide use of a $17 million, two-way communications-software system developed for SmartGridCity.

“Public Service, when it steps back, has admitted that it wasn’t handled the way they thought it should have been,” said William Levis, director of the Colorado Office of Consumer Counsel.

“But they are still arguing they are entitled to their money,” Levis said.

Xcel’s 1.4 million Colorado customers are already paying for $27.9 million in SmartGridCity costs, under a 2011 PUC ruling.

An administrative-law judge’s recommended decision on whether to grant the last $16.6 million is expected in the next few weeks and will then go to the full commission.

The project has provided tools for better managing electricity and identifying failing equipment, the company said.

“It delivered on this package, and that is what we are asking customers to pay for,” Hyde said.

The project’s impact will spread beyond Boulder, said David Eves, Public Service’s president and chief executive.

“We invest about $200 million a year in our electric-distribution system,” Eves said. “SmartGridCity is providing valuable insight to inform our investment decisions.”

Former Xcel CEO Dick Kelly, who approved SmartGridCity and retired in August 2011, did not return calls for comment.

•••••

In the past few years, utilities nationwide have embarked on smart-grid projects that combine enhanced communication through two-way, or “smart,” meters with sensors, software and computers.

Executives for Minneapolis-based Xcel, which operates in eight states, went looking for a site in 2007. The choice came down to Boulder or St. Cloud, Minn.

Boulder officials were intrigued.

“We thought it was an excellent opportunity to work with Xcel on something that would revolutionize how we consume energy,” said Shaun McGrath, Boulder’s mayor at the time.

SmartGridCity would deploy thousands of sensors in the distribution system and 35,000 smart meters that would enable as many as 10,000 homes to communicate with the utility, right down to thermostats and appliances, according to Xcel presentations.

McGrath said he was told by Xcel officials that the company would not seek to have customers pay.

Xcel said the company meant it would not seek immediate recovery or charge Boulder residents.

At the time, Boulder had a task force and a consultant exploring whether to renew Xcel’s 20-year electricity franchise as the city’s power supplier or to replace it with a municipal utility.

“The franchise was an issue,” said former Boulder City Manager Frank Bruno. “They didn’t want to make an investment in a city they didn’t have a relationship with.”

In March 2008, Bruno recommended shelving work on a municipal utility and moving forward with SmartGridCity. The City Council agreed.

Xcel had assembled seven consortium partners — high-tech and engineering companies — that would participate in the project and share in the cost.

SmartGridCity was to be a $100 million project, Xcel said, but the utility’s share would be $15.3 million, with consortium partners picking up the rest.

Xcel executives presented their vision for SmartGridCity to the PUC in May 2008. Not included was a detail discussed a week earlier at an internal Xcel corporate-finance meeting: Xcel’s share of the project’s cost already appeared to have risen to as much as $31 million.

In June, Xcel’s board approved SmartGridCity with a project limit of $25 million, according to a company document.

“We underestimated the cost of construction,” Xcel’s Hyde said of the rising price. “We also assumed some costs would be assumed by consortium members that were not.”

The communications technology that Xcel chose for the project was broadband over power line, or BPL. Other smart-grid pilots were using Internet or other wireless communications.

Xcel opted for BPL because “it was faster than anything else,” said Randy Huston, Xcel’s director of infrastructure and smart grid. “It was intended to be a test bed for 20 years.”

BPL, however, has limited range, so Xcel also installed a fiber-optic network. The costs of laying the lines in Boulder’s rocky soil soared 74 percent over initial estimates to $21 million.

BPL is working in Boulder, but the technology, which Xcel describes as reliable but expensive, will not be deployed elsewhere, Huston said.

“BPL is an orphan technology,” said Schoechle, who serves on a smart-grid working group of the International Electrotechnical Commission.”No one in the U.S. is using it.”

•••••

Through the summer of 2008, crews worked to get the system running in time for the Democratic National Convention in Denver that August, company documents show.

On the convention’s second day, Xcel unveiled the first SmartGridCity home — with integrated software, solar panels and an electric-car charger — and gave tours, including one to ABC’s “Good Morning America.”

As 2009 began, costs continued to mount and some partners pulled back from the project. That’s when Xcel officials realized the utility’s share of project costs would rise, the company said.

“You’ve got to remember we were going into a recession right at that time,” said Xcel’s Hyde. “There was a lot of turmoil.”

In March, an executive leadership review laid out three options: stopping deployment with the 14,653 smart meters installed, modifying the program or completing the project with 23,000 meters. Xcel chose to proceed with the full project.

“Management concluded the costs of continuing forward with the project completion under the initial scope were less than the potential benefits that could be determined if there were a broader pilot,” Scott Wilensky, an Xcel senior vice president, said in PUC testimony.

One issue was whether scaling back the project would put at risk Xcel’s ability to keep the Boulder franchise.

In April, a presentation to the board outlined the remaining work and set the budget at $27.3 million.

That spring, the two executives in charge of SmartGridCity left Xcel.

Ray Gogel, the lead executive on SmartGridCity, became CEO at Current Group, one of the consortium partners. Mike Carlson left Xcel for a position with another consortium partner, GridPoint Inc.

Gogel did not return calls for comment. Carlson did not reply to e-mail requests for comment.

In May, Xcel filed for a $132.5 million rate increase for Colorado that included $27 million for SmartGridCity.

The project was being reviewed at top levels of the company, but Hyde said it “didn’t have as rigorous a process around it as our normal controls.”

Wilson Gonzalez, an energy consultant and expert witness for the consumer counsel’s office, said, “SmartGridCity had no cost-benefit analysis, no business plan, and without those, you don’t know if costs are legitimate.”

•••••

Soon after the rate case was filed, a decision was made in Minneapolis to move nearly $17 million in SmartGridCity-related software costs from Xcel’s normal information-technology budget into the project.

Xcel had assumed that software to enable two-way meter communication would be used systemwide, but the company abandoned that idea, Hyde said. In July 2009, Xcel revised its rate request, seeking $42 million for SmartGridCity.

Xcel’s inclusion of SmartGridCity in its 2009 rate case sparked challenges that the project had not been properly approved by the PUC.

A new project usually needs to obtain PUC approval before it starts along with an approved budget.

Xcel argued that it didn’t need approval or a special budget review because the project was just an “upgrade” of Boulder’s local power lines, but the commission didn’t agree.

“You don’t have ‘only improvements’ of a distribution system discussed on ‘Good Morning America,’ as this was,” Ronald Binz, then-chairman at the PUC, said at one hearing.

SmartGridCity’s price tag, now $45.8 million including interest costs, was challenged by the consumer counsel’s office — which represents residential customers and small businesses — and industrial customers Climax Molybdenum and Rocky Mountain Steel.

Harry DiDomenico, a PUC analyst, testified that the project “was conducted outside of normal budgeting processes and was therefore never subject to normal budget reviews, policies and internal audit procedures.”

The consumer counsel’s office said that once the project hit $27 million, Xcel should have stopped. Any costs beyond that were imprudent.

Climax Molybdenum and Rocky Mountain Steel in filings argued that SmartGridCity was a research project, that it had not achieved its goals and that shareholders, not ratepayers, should pick up the bill.

The PUC gave Xcel $27.9 million in rates for SmartGridCity in January 2011 and said the company could get the remainder if it showed value to customers and had a strategic plan for the project.

The commission expressed concern over whether SmartGridCity would become “an integral part of the distribution system.”

“I’m not suggesting another dime could not be spent on SmartGridCity,” said commission chairman Binz. “We just want to know what is going on.”

In November, Boulder voters, frustrated that Xcel wasn’t supplying more renewable energy, passed a local tax that enabled the City Council to begin exploring whether it could replace Xcel with a municipal utility.

Xcel had just begun an in-home pilot using a $700 package from Brooklyn, N.Y.-based EnergyHub. The utility was forced to find a new supplier after partner GridPoint stopped making in-home devices for Xcel and another vendor exited the business.

Xcel had set a goal of installing 1,850 units in Boulder. A total of 101 Boulder homes received them in 2011.

The utility has also distributed 1,000 in-home units in Westminster and Centennial — saying it wants to test the technology in a variety of settings.

The company can do that because the units communicate using the Internet, not the smart grid. The smart meters Xcel installed on Boulder homes, replacing standard electricity meters, aren’t capable of communicating without upgrades.

“So at the end of the day, SmartGridCity was irrelevant,” said smart-grid analyst Schoechle.

•••••

Xcel is now seeking the last $16.6 million, saying the system is up and running and has provided valuable insights.

The utility has found that through the smart grid it can better manage voltage, saving fuel and cutting customer complaints about fluctuations to zero.

Adding so many sensors also has given Xcel engineers a better idea of the limits and vulnerabilities of the system, Xcel’s Huston said.

While grid-monitoring sensors won’t be placed across the entire Xcel network, they can be deployed in areas with older equipment or congestion problems, Huston said.

Lessons learned from SmartGridCity also are being used in the company’s network strategy plan for IT investments.

Boulder officials question that conclusion. “There hasn’t been any substantial cost-benefit analysis,” said Kelly Crandall, a city sustainability specialist. “You want to encourage innovation, but this is not a good model.”

The future of SmartGridCity is uncertain. The utility is running two pilots in Boulder — the in-home device test and a pricing test that offers lower charges for managing electricity use during peak periods. Any future pilots would wait until those are completed at the end of 2013, Hyde said.

Because of depreciation and delays in getting SmartGridCity investments into rates, she said, Xcel will recover just a fraction of the money it has spent, even if it gets the $16.6 million.

In explaining SmartGridCity, Xcel’s Wilensky said in testimony that the technologies were “new and not ones the company had experience with” and that “the project wasn’t nearly as well defined as a typical project.”

That was SmartGridCity’s problem, said Wade Troxell, a Colorado State University engineering professor.

“SmartGridCity was more marketing than designed to be an effective system,” said Troxell, an emeritus member of the Gridwise Architecture Council. “It didn’t have a design and a definite purpose. It isn’t surprising that there were cost overruns.”

Mark Jaffe: 303-954-1912, mjaffe@denverpost.com or twitter.com/bymarkjaffe

 


 

 

A look at Xcel’s SmartGridCity

 

2008

Jan. 16: Xcel Energy announces that it is joining with seven high-tech and engineering firms to build a project called SmartGridCity. Xcel’s share is to be $15.3 million.

March 12: Xcel announces Boulder as the site for SmartGridCity.

May 15: Xcel begins work on SGC.

May 22: The Xcel finance committee is told the cost of the project to Xcel could reach $31 million.

November: Xcel files with the Colorado Public Utilities Commission, saying the cost of SGC is $15.3 million.

2009

April: Xcel board is updated. SGC is reclassified as “a major capital project” with a price tag of $27.3 million.

May 1: Xcel applies for a $132 million rate increase that includes $27.3 million for SGC.

July 10: Xcel revises its rate-increase application to $180 million, including $42 million for SGC.

Sept. 8: Xcel says, “SmartGridCity becomes first fully functioning smart city in the world.”

2010

March 10: Xcel files for commission review and recovery of $44.5 million in costs.

Oct. 1: Xcel begins an SGC pricing pilot that includes about 4,000 homes.

2011

Jan. 5:PUC approves SmartGridCity and allows Xcel to add $27.9 million to customer rates but withholds $16.6 million until Xcel can show customer benefits.

Oct. 3:Xcel begins in-home device pilot in Boulder, and also in Westminster and Centennial. Set to finish in 2013.

Nov. 11: Boulder voters approve ballot measures authorizing the City Council to explore replacing Xcel with a municipal power utility.

Dec. 14: Xcel files to recover the remaining $16.6 million.