The safety of the STARS program is related in obvious fashion to its reliability. No one denies that if a STARS rocket fails, public health and the environment are far more likely to be harmed than if the rocket is launched successfully.
When it comes to reliability, however, the STARS has no track record to judge. Until the February launch, its components had never been tested as a unit. Even with that one success, the jury is still out on the rocket’s overall reliability.
Classified Information
The government’s environmental impact statement claims that the STARS rocket is 97 percent reliable. The basis of the claim is a memo prepared in 1990 by two scientists with Sandia National Laboratories, K.C. Abbott and R.W. Plowman.
Abbott and Plowman write: “The mission reliability prediction … in this report is unclassified, since the reliability values for the Polaris stage 1 and stage 2 motors are not included. Also, reliability data for a portion of the Polaris Thrust Vector Control (TVC) subsystem has not been included … Thus the reliability prediction must be seen as a best-case prediction, assuming the reliability of the excluded components is 1.0. The classified annex contains the classified mission reliability prediction.”
Abbott and Plowman take note of the aging problems in the first-stage boosters that cause the insulation to deteriorate. The refurbished first-stage booster was successfully tested in just one static test firing. Aging problems in the second-stage booster were described as well.
In attempting to develop an accurate prediction, Abbott and Plowman undertook a data search to find out what information existed on Polaris reliability. They report having found a data book, which contained prediction information. In their unclassified memo, however, no specific mention is made of any Polaris reliability record or prediction. Nor does the EIS mention any classified report or memo that might give a more realistic figure for the STARS reliability prediction.
A Crap Shoot
The Sierra Club Legal Defense Fund sought out David Wright, senior staff scientist with the Union of Concerned Scientists in Boston, to look into the matter of the reliability of STARS rockets. Wright, who was awarded M.S. and Ph.D. degrees in physics from Cornell University, reviewed the Abbott and Plowman report and found it worthless as a basis for making public policy.
The method of Abbott and Plowman is “fairly standard,” Wright says in his comments, but goes on to note that the authors themselves did not intend for their calculations to support “a reliability figure of 97 percent for the STARS vehicle.” They admit that their analysis is “a prediction, not an assessment” of STARS’ reliability, Wright says, “meaning that it relies on engineering estimates or ‘typical’ reliability figures for missile components but is not based on data directly applicable to the STARS booster; thus its applicability to the particular case of the STARS system is unknown, and cannot be known in the absence of component testing and integrated flight testing of the STARS vehicle.”
Wright examined actual flight data for the first and second stages of the Polaris motors, arriving at an overall reliability rate of 95 percent (although most of the flights that constitute that data base were made by Polaris during the peak of the Polaris program, before the aging problems became manifest). After integrating that with the data used by Abbott and Plowman in predicting the reliability of other STARS elements, Wright estimates an overall STARS reliability of 92.8 percent.
If this 92.8 percent figure is borne out in actual flight experience, Wright says, “there will be a greater than 50 percent probability of at least one failure within 10 launches.
“Assuming a test schedule of four launches per year, this means that there would be a greater than 50 percent chance of seeing a failure by the third year of STARS launches.”
To try to arrive at better model, Wright analyzed the launch records of a similar missile system, the refurbished Minuteman I, and found it had a reliability record of 82 percent.
If this reliability rate — 82 percent — were duplicated in the STARS system, an assumption that Wright believes is reasonable, “there would be seven expected failures in the full launch series,” he concludes.
Army: Trust Us
Joining Wright in his dismay over the state’s use of the high reliability prediction in a policy-setting document was Robert Aldridge, an aerospace engineer who helped design the Polaris A3 rockets. In an affidavit, Aldridge noted that the Sandia memo “appears to be an internal prediction of only the newly designed components for the STARS. The authors have warned that it has limited value…. Consequently, the reliability predicted in the memo should not have been used in the DEIS or EIS to assess the impact on public health and safety or on the environment, and should not have been relied upon by the state in its EA” for the memorandum of agreement.
The Sierra Club Legal Defense Fund brought out the issue of overstated reliability figures in several memoranda filed with the state Circuit Court over the adequacy of the state’s environmental assessment. The state’s response was to argue that because Judge Ezra did not rule that it was an issue in the federal EIS, the Sierra Club Legal Defense Fund “should not be allowed to relitigate in this forum of second choice” the issues of safety and reliability.1 The state made no effort at all to defend the 97 percent reliability prediction.
In an amicus memorandum filed with the state court, the U.S. Army takes up the matter of reliability predictions. Begging the obvious, the Army lawyers note that “Sierra Club neglects to mention that the actual reliability data were included in the report’s classified annex, which was available to deciding Army officials but was not released to the public.”
Finally, the Army notes that even if Wright’s prediction of a 92.8 percent overall missile reliability figure were true, “the original design reliability goal of the Strategic Target System of 90 percent is still met.” In other words, out of the 40 STARS missile launches planned, as many as four of them are predicted to fail, and the Army would still claim success for the project inasmuch as its design goals were met.
1 The federal court decision by Judge Ezra had been vacated, however, before it could be heard on appeal. As such, it carries no legal weight. In any case, the ruling of Judge Fong in July 1992 was that the state’s compliance with state statutes was properly a matter for state court. If it was a “forum of second choice” at all, it was so for the state, it would seem, which had sought to remove the case to the federal level.
Volume 3, Number 9 March 1993
Leave a Reply