The cost problems that are expected to force the restructuring of a program to develop a new constellation of U.S. government weather satellites can be largely traced back to assumptions made early on in the program that a key sensor would be relatively easy to develop, according to industry and Pentagon officials.

Those assumptions later proved to be false, and the size of the resulting cost overruns required the Pentagon to notify Congress Sept. 28 that the price of the National Polar-orbiting Operational Environmental Satellite System (NPOESS) will be at least 15 percent higher than the Defense Department’s current estimate of $6.8 billion.

The Pentagon was forced to notify Congress about the cost growth on NPOESS due to legislation known as the Nunn-McCurdy provision.

The NPOESS program’s cost growth is forcing the government to look at options for the program that include delaying the launch of the first satellite four years later than the current date of 2010; eliminating certain instruments from the first two satellites; and dropping plans to launch an experimental spacecraft that would test out key NPOESS instruments before the launch of the first operational satellite.

The NPOESS system, which is being integrated by prime contractor Northrop Grumman Corp., is intended to provide weather data to military and civil customers, and is jointly funded by the U.S. Air Force and the National Oceanic and Atmospheric Administration.

Lt. Col. Karen Finn, a spokeswoman for the Air Force, said in a written statement that the development of the sensors for the NPOESS satellites has proven to be more difficult than expected. While some of the key sensors — most notably the Visible Infrared Imager Radiometer Suite (VIIRS) — are based on those that already are in use in space today, there have been issues with “scaling up the sensors” for NPOESS, she said.

“It is the combination of system engineering, design, management, fabrication, integration and test, quality processes and inadequate levels of employee experience that contributed to cost and schedule overruns with VIIRS, making the current program baseline unexecutable,” Finn said.

The VIIRS instrument, which is intended to provide imagery, sea-surface temperature readings and ocean color measurements, is built by Raytheon Co. of Santa Barbara, Calif.

The government program office and Raytheon initially believed that the VIIRS instrument would be a “close cousin” to the Moderate Resolution Imaging Spectroradiometer that Raytheon had built for NASA’s Aqua and Terra Earth observation satellites, and that only minor technical changes would be necessary, industry sources said.

Northrop Grumman was not involved at that point, because the government awarded the instrument development contracts on the NPOESS program prior to the award of a contract to Northrop Grumman in 2002 to be the lead systems integrator for the program .

Based on this assumption, the program office allowed Raytheon to pass an early milestone, known as the critical design review, without building a prototype sensor, the sources said. Prototype hardware called an engineering development unit is normally used to evaluate hardware design at the critical design review phase, but in this case the development unit was not built until later, the sources said.

This built a shaky foundation for the sensor program, and program officials today are still dealing with the “lingering efforts of a poor design,” the sources said.

The Moderate Resolution Imaging Spectroradiometer design has since proven to be of little use on the VIIRS effort, the sources said.

John Leslie, a spokesman for the National Oceanic and Atmospheric Administration, said that a partial engineering and development unit was built prior to the critical design review. Program officials had felt that a full prototype of the VIIRS sensor would not be needed at that point based on the expected similarity to the Moderate Resolution Imaging Spectroradiometer, Leslie said.

The industry sources said that the VIIRS sensor required focal planes that were so different from those of earlier instrument that engineers had to start from scratch. As engineers worked on the VIIRS engineering development unit in mid-2003, they found information from sensor cells leaking into other cells, causing a problem called “cross-talk” that would damage the data provided from the instrument, the sources said.

Engineers were able to fix the interference problem, but by mid-2004, encountered further difficulty when parts started breaking during vibration testing of the engineering development unit, the sources said.

At that point, officials from Northrop Grumman and the government program office began reviewing VIIRS to determine whether the problems with the design were isolated or widespread.

When hardware intended to cool the instrument’s focal plane failed during testing of the engineering development unit in November, program officials knew that the cost and time needed to redesign various VIIRS components would significantly disrupt the program’s schedule.

Program officials found that one of the roots of the problems with VIIRS was the quality of personnel working on the effort, the sources said. Many of the people that had worked on the Moderate Resolution Imaging Spectroradiometer did not join the VIIRS program, and Northrop Grumman and Raytheon each brought in experts to help get the sensor work back on track, the sources said.

More than 80 percent of the cost growth on the program is due to instrument problems, with VIIRS as the most glaring example, but not the sole culprit, the industry sources said.

Sources interviewed for this article emphasized that the NPOESS problems have been with the individual instruments, rather than integrating too many sensors onto the spacecraft, requirements creep or the recent decision to add a wide-area imagery payload.

Sally Koris, a spokeswoman for Northrop Grumman Corp., said in a written statement that Northrop Grumman is taking the NPOESS issues “very seriously” and is working with the government to limit the rising cost.

“We have responded to problems with aggressive action to mitigate the impact of cost and schedule issues, including actions aimed at correcting systemic problems with selected subcontractors,” Koris said. “Northrop Grumman and its teammates are completely dedicated to achieving mission success on this critical program to ensure this much-needed capability is delivered to military and civilian users.”

Raytheon spokeswoman Sabrina Steele deferred to Northrop Grumman for comment.

In contrast to the story with VIIRS, some of the instrument work is going relatively well, the industry sources said. Officials working on the Cross Track Infrared Sounder, which is built by ITT Industries of Fort Wayne, Ind., have produced several engineering development units and have encountered only minimal problems due to their heavy attention early in the program, the sources said.