Automating NIF Beam-Timing Delays

For many NIF experiments, data from different stages of the experiment as it evolves in time can provide insights into such areas of interest as the ability of materials to resist growing hydrodynamic instabilities and the properties of low-density plasmas as they expand away from the target.

To increase the facility’s usefulness to experimenters, in 2013 the NIF Team developed a unique long-delay backlighter (LDBL) capability that made it possible to propagate backlighter beams at delays of up to one microsecond after the initial laser pulse (backlighters irradiated by several of the NIF beams provide the illuminating radiation that enables diagnostic equipment to record experimental results). The long-delay capability has been used in more than a dozen NIF material strength and other high energy density experiments in the last two years (see "Experiments Test High-Pressure Material Response").

But setting up an experiment for an LDBL shot, and then restoring the system to normal operations, was time-consuming and labor-intensive. "The overhead was quite extensive," said Dave Mathisen, Integrated Computer Control System shot timing system engineer. "There was always at least a four-hour delay in just setting up for one of these shots."

The shots required the installation of additional lengths of fiber in the Master Oscillator Room (MOR), where each NIF shot originates, to shift the timing of the laser "window" on the backlighter beams (see "Injection Laser System"). Then a software script was run to redefine timing points throughout the beamline for the value of the additional fiber, while saving the initial values. "It would basically be tricking the shot automation system downstream to not really know whether or not we had a (new) fiber installed at all," Mathisen said.

"Then there was a manual inspection of hundreds of lines of data," added Gordon Brunton of NIF Computing Applications. "Changes that were being proposed needed to be reviewed by subject-matter experts to make sure that what we were intending to do was going to achieve the desired results." After the shot, the extra fiber was uninstalled, the set points were restored to their original values, and the normal timing was verified. The total process required about four hours before the shot and another four hours afterward, or about one complete shot cycle.

This year, as part of NIF’s continuing efforts to increase its shot rate in part through efficiency improvements, a multi-program team set out to streamline and automate the LDBL process. "We took a long view of it," Mathisen said. "We met with customers, we met with the User Office, we met with everybody who had a stake in it." Shot RIs (responsible individuals), fiber fabricators, experimentalists, diagnostics analysists, and shot automation and logistics experts were consulted to develop a new process, called the fiber delay backlighter (FDBL) system.

Set Procedures and Guidance

"It was an automation of the process on the front end," Mathisen said. When users decide to employ the long-delay process, the system provides feedback about the availability of the required fibers and the cost of ordering new fibers, if needed. If the user approves, the fibers, which are fabricated in-house, are prepared and the experiment is run.

The system also provides set procedures and guidance that allows the fibers to be installed in the MOR by trained technicians, eliminating the need for timing experts to handle the setup. Data system tools validate that the timing system is capable of supporting the requested delays. "That was all manually done previously by manual timing qualification," said Brunton. With the FDBL process, setup and teardown time has gone from four hours before and after the shot to 40 to 50 minutes per shot, depending on the number of quads employing the delays.

Another FDBL feature is the ability to change the delay time in different quads, so data from multiple backlighters can be obtained. "The shot users are all very excited about using the system because it’s much easier for them to visualize," Mathisen said. "They get instant feedback when they’re planning. All kinds of optimizations fall out of this because of the ability to see what we’re doing."

The FDBL system was used in four recent Discovery Science experiments designed to study the star-formation process in molecular hydrogen clouds, such as the famous "Pillars of Creation" in the Eagle Nebula (see "Unlocking the Secrets of Star Creation"). Researchers were able to use FDBL to drive four hohlraums one after another, from times of 0 to 15 nanoseconds, 15 to 30 ns, 30 to 45 ns, and 45 to 60 ns, for a total x-ray drive length of 60 ns. The multi-hohlraum array simulates a bright, sustained stellar source, and the NIF Eagle science package mocks up a radiatively-driven, star-forming cloud of molecular hydrogen. The same fiber was used in back-to-back experiments, saving additional time.

"The fiber-delay system permits the Eagle multi-hohlraum x-ray source to drive a science package for 60 nanoseconds, instead of the 30 nanoseconds that was possible on the first NIF Eagle shots" in April, said co-principal investigator Jave Kane. "This permits the science package to evolve hydrodynamically into a mature column structure with density and velocity similar to what astronomers observe in the Pillars of the Eagle Nebula."

"Now that we’re allocating shots based on time, (the FDBL process) empowers the users," Brunton said. "If they can avoid a fiber swap, they can avoid the time penalty cost in doing that by switching the order of the shots in the experiment. And this is even more important going forward into (Fiscal Year 2016). The shot rate (goal) goes up to 400 a year now (from 300 in FY 2015), so anything we can do to try to minimize the impact of setting up for these more complex experiments can pay dividends."

Other members of the NIF Team contributing to the development of the FDBL process were Adrian Barnes, Mitanu Paul, Bruce Conrad, Gaylen Erbert, Allan Casey, Russell Fleming, Rich Beeler, Rick Olson, Mark Bowers, Mike Shaw, Ron House, Barry Fishler, Joyce Li, Steve Hahn, Scott Reisdorf, Hye-Sook Park, Bruce Wilson, Susheela Muralidhar, Yvon Tang, Mike Shaw, Doug Speck, Misha Shor, Brent McHale, Steven Glenn, Susan West, Yan Pan, and Brett Raymond.