As the nail pierced one of the eight cells, it accomplished its goal of setting off a short circuit. But only smoke, not fire, belched from the battery.

But the test and other evaluations that Boeing conducted while the plane was in development proved to be far off the mark in predicting what would happen when the plane was in use and led Boeing to severely underestimate the likelihood of a fire in the 63-pound battery.

Now, as Boeing tries to recover from a battery fire on one of the jets and a smoke incident on another that led to a worldwide grounding of the 787s in January, federal officials say it is crucial that the company collect better data this time from its laboratory and flight tests. That is particularly important, they said, since the cause of the jet’s battery problems is still unknown.

Boeing won approval on Tuesday from the Federal Aviation Administration to begin an extensive array of 20 types of tests to determine if its proposed new battery protection system works.

“It is clear from what happened in January that there are other ways for the batteries to catch fire that clearly weren’t captured by the nail test,” said Donald R. Sadoway, a materials chemistry professor at M.I.T.

He said the test — it is standard in testing lithium-ion batteries in cellphones, laptops and electric cars — measures just one kind of failure mode and represents an extreme short circuit of the battery in a very short period of time.

“It can work on a cellphone battery with a single cell, but you’re talking about a very different configuration here,” he added.

A preliminary report on the fire that occurred in a plane parked at Logan Airport in Boston on Jan. 7, released last week by the National Transportation Safety Board, provided new details of Boeing’s initial battery testing, including its reliance on the nail test for some of the basic data on the odds of incidents with smoke and fire.

The board’s report said that Boeing also gathered information from companies using the lithium-ion batteries in other industries and, based on that and the test data, it calculated that the odds of smoke being emitted from a battery was only one in 10 million flight hours.

Given that Boeing has never made its own batteries, it assigned other crucial tests to its subcontractors, including GS Yuasa, the Japanese battery maker, and Thales, a French company that oversees the battery charging system.

The F.A.A. also delegated much of the testing to Boeing, except for a safety assessment conducted once the plane’s electrical power system was certified. The F.A.A.’s decision to delegate some of its certification authority was done under a decades-old program meant to help the agency keep up with fast-changing technology but also to fill a gap in its own expertise.

In the course of the original testing, the batteries were also subjected to other kinds of destructive tests, including provoking an external short circuit, overcharging the batteries for 25 hours, subjecting them to high temperatures of 185 degrees Fahrenheit for an extended period, and discharging them completely. Boeing found that none of these tests in its computer modeling resulted in a battery fire, according to the safety board’s report, leading it to conclude that the chance of a battery fire was one in 1 billion flight hours, a negligible probability. In fact, the Boston fire occurred after the entire 787 fleet had flown with passengers for only 52,000 hours.

F.A.A. and Boeing officials said they did the best they could, given what was known about the batteries when the agency approved them for the 787 in October 2007. The innovative plane, which uses lighter-weight composite materials to cut fuel costs by 20 percent, makes more extensive use of the new batteries than any other commercial aircraft, and the F.A.A. imposed a series of special safety conditions to compensate for that risk.

Article source: http://www.nytimes.com/2013/03/14/business/initial-tests-of-battery-by-boeing-fell-short.html?partner=rss&emc=rss