Tracking data on PG&E's 12,000 volt powerline tied to deadly Zogg fire reflected repeated anomalies leading up to the fire. But, the problems were sporadic and didn’t last long enough to trigger an automatic shutdown under PG&E’s equipment settings, according to the company federal court filing Monday.
The fire started Sept. 27 in Shasta County - killing four people, charring 56,000 acres and destroying more than 200 structures.
While the cause is still under investigation, PG&E confirmed Monday that Cal Fire has seized a burned cross arm on one of its poles, shattered insulators and portions of a Grey Pine growing about 60 feet from the powerline - known as the Girvan distribution circuit.
It is not clear what role any those pieces had in sparking the fire.
According the company’s filing with U.S. Judge William Alsup, who is overseeing PG&E’s felony probation stemming from the San Bruno gas explosion case, the Girvan Circuit was patrolled in the spring of this year - and many trees were trimmed or cleared on Zogg Mine road. However, the company was not able to say whether the specific tree seized by Cal Fire was one of them.
The company said conditions did not merit cutting power to that specific distribution line as part of the ongoing preemptive power shutoff that day.
Power shutoffs – also known as de-energizations -- are typically ordered in cases where they expect sustained winds of more than 25 mph and gusts of more than 45 mph. The sustained winds predicted were only 15 mph that day, with gusts at about 32 mph.
“The readings from all three stations were below PG&E’s general de-energization thresholds,” the company said. “They also were below the general de-energization thresholds in place during the 2019 fire season.”
In its chronology, the company said its records show anomalies first showed up in its automated control system at 2:40 p.m., with a drop in voltage recorded on an electronic meter near the fire's ignition point.
Over the next several minutes, several more power fluctuations occurred. While such fluctuations can point to problems, PG&E told Alsup that the fluctuations didn’t last long enough to trigger the devices that could automatically shut down lines.
For the circuit breaker devices to be triggered at that location the fluctuation had to last for at least 20 to 25 seconds. If the duration is shorter, the company’s system ignores it, to guard against unnecessary power outages.
“The delay avoids the operation of the protective devices in response to transient conditions, such as normal changes in loads on the line,” the utility said in its filing to the judge.
That meant despite the trouble signs, the line was still energized at 2:42 p.m., when smoke was first spotted on a fire-watch camera in the area. A minute later, three electronic meters reported having lost power - another sign of trouble.
Then, at 2:43, a weather satellite picked up a sizable heat signature in the same area, north of Igo, but did not record that heat as being from a fire.
PG&E’s system, meanwhile, reported more power fluctuations - but they too weren’t long enough to trigger an automatic shutdown.
By 2:46, the satellite data reflected a fire, and 14 minutes later, a PG&E technician spotted the flames. However, it was not until 3:06 p.m. that a power fluctuation lasted long enough to trigger line shut down, according to PG&E’s filing.
Dan Mulkey, an engineering consultant who previously worked for PG&E, said the company’s equipment settings are more sensitive than in the past, but that means system operators can be flooded with reports and have to wait for technicians to diagnose them.
“The lower you set it, the more likely you are to lose power under normal conditions,” he said. “It starts to be a nuisance trip. You shut it down for no reason, when there’s nothing wrong.”