Page 22


FEBRUARY 20, 1987 struction manager of the project and assigned the San Francisco based Bechtel Corp. to finish the job. Brown & Root voluntarily withdrew as builder and was replaced by Ebasco, the company that should have been on the job in the first place. Never before had an American nuclear plant been forced to change engineer and builder during construction. Originally projected to cost less than a billion dollars, the STP is now seven years behind schedule and budgeted at $5.5 billion. The bottom line would have been even worse if the NRC had not blown the whistle on Brown & Root in 1980. The four STP partners subsequently sued Brown & Root and its parent company for mismanagement, and Brown & Root reluctantly agreed to pay the partners $750 million in damages. Since Comanche Peak was being built on a parallel path to the STP’s, it would have been prudent for Texas Utilities to put Brown & Root’s performance under the most severe scrutiny. But it didn’t. Apparently neither Texas Utilities nor Brown & Root learned muchfrom the South Texas experience. They just blundered on. Construction was begun on both the STP and Comanche Peak plants before even 20 percent of the engineering was completed. At Comanche Peak, the engineering, done by Gibbs & Hill of New York, never got up to speed. In an excellent package of stories on Comanche Peak published Jan. 4, the Fort Worth Star-Telegram reported, “Former plant workers said that because the design process was so cumbersome and often unrelated to true conditions at the site, construction crews sometimes ended up doing the designing. When a craftsman set out to do a construction task, design documents that acted as his instructions might have included as many as 32 revisions of a pipe or a pipe hanger or a cable tray installation.” The Star-Telegram continued, “Often . . . the craftsman would locate the place where the piece of equipment was to be installed, only to find another pipe or duct or cable tray in the way. Instead of sending the problem back through the engineering department to be solved properly, the former workers say, the welder or pipefitter would decide how to fix it. After the work was done, the new drawings would be sent back to engineering to be approved after the fact. . . . Analyses that would check whether the alteration would actually work under stress were left undone.” In 1984, Texas Utilities believed that the plant was substantially complete and asked the Atomic Safety and Licensing Board, which is part of the NRC, for an operating permit. At the time, the major criticism of the project had come from citizen intervenors who had learned of safety hazards from plant workers and inspectors. Brown & Root, Gibbs & Hill, Texas Utilities, and, astonishingly, top administrators of the NRC’s Arlington regional office, all had ignored the rumblings of trouble. Now the citizen intervenors commanded the attention of the NRC’s top brass. In response to allegations of cavalier indifference toafety regulations, the NRC assigned a special team of 75 consultants and investigators to reinspect the plant. ANUCLEAR PLANT proves it is safe to operate through a paper trail. Each piece of pipe, each batch of concrete associated with the plant’s safety system must have its own set of records certifying how and when it was made and where it was put in compliance with the engineering plan. Without proper documentation, a plant might as well have not been built, because it cannot be certified as safe and therefore cannot be allowed to operate. There are 11 million pages of paper on every aspect of construction of Comanche Peak. But the investigators discovered that what is on the ground and in the files is sometimes at odds. Safety reports were altered and inspectors harrassed to give Comanche Peak passing grades. This was done not only by Brown & Root workers but also by senior officials in the NRC Arlington regional office. In light of these revelations, the NRC decided that it could not accept the tainted safety records compiled by either the builders of Comanche Peak or the NRC Region IV office in Arlington. That is why Comanche Peak is now undergoing a meticulous reinspection overseen by a newly created special projects office in Washington, D.C. Texas Utilities predicts that the reinspection will be successfully completed in 1988 at an additional cost of $2 billion, bringing the final tab for the project to $7.7 billion, ten times its original estimate. Interest payments alone on Comanche Peak’s construction loans are running $1 million a day. Texas Utilities now alleges that the plant will begin operation in 1989. But many knowledgeable observers believe that the NRC will force Texas Utilities to mothball the plant. There are precedents: the Zimmer nuclear power plant in Moscow, Ohio, was abandoned when it was 97 percent complete because safety could not be assured. Similar