BY JIM ALTENBERG
MARTINEZ, California - A report issued by the Tosco Refining Co. on the Jan. 21, 1997, explosion at its Avon Refinery attempts to place responsibility for the disaster on the seven operators who were working in the unit at the time. The blast took the life of Michael Glanzman, a hydrocracker operator. Forty-six others were reported injured, including maintenance workers, workers employed by contractors in the plant, and supervisors. Following investigations by the Contra Costa County Health Services Department, the California Occupational Safety and Health Administration, and other government agencies, as well as Tosco itself, two reports were presented to a meeting of the Contra Costa County Board of Supervisors June 3. One was issued by the company, and the other by Health Services.

While few workers at Avon have actually read the reports, a widely distributed company flyer purporting to summarize the company's findings was met with anger and derision. Many were not surprised when the company tried to pin responsibility for the blast on the operators. A cover letter to the Tosco report began with pronouncements of the company's concern for safety and regret for the tragedy of January 21. It emphasized the report's "comprehensive, very detailed, and by its nature, very technical" aspect, in order to discourage anyone from studying it, and to bolster the image of an objective analysis of the disaster.

Criticism of report
Officials of the Oil, Chemical and Atomic Workers union (OCAW) also participated in the investigation. But in the end, they refused to endorse the report's conclusion, charging that Tosco said little about the company's unwillingness to require that workers adhere to established procedures when faced with potential runaway reactions. Union officials also said the OCAW Health and Safety Representative in the plant had been excluded from investigative meetings where the company management's role was taken up.

The County Health Services Department sent a letter to the Contra Costa Times protesting the paper's May 30 headline "Workers erred in fatal blast at Tosco." This "draws one to a conclusionary summary of our report that is not accurate," the letter stated. Workers posted up the Health Services letter in control rooms and maintenance shops throughout the Tosco plant.

The Communities for A Better Environment (CBE), a San Francisco-based environmental organization that has spoken out against oil refinery accidents and pollution, also responded to the report. The CBE said Tosco management should have shut down the hydrocracker rather than run with unreliable monitors and leaking equipment. Faced with an increasing number of refinery accidents and leaks, residents of towns near Contra Costa County's four refineries have demanded more stringent laws governing refinery operations and land use. They have also set up their own air monitoring systems to collect independent data on refinery emissions, spills, leaks, and accidents.

Hydrocrackers produce large amounts of high-profit gasoline and diesel fuel from the heavier materials found in crude oil. This is accomplished by combining hydrogen with gas-oil or diesel in catalytic reactors at pressures as high as 1,700 pounds per square inch (psi).

The reports explain that plant procedures require operators to shut down and depressurize the hydrocracker when reactor temperatures reach 800 degrees, lest a dangerous temperature runaway or "excursion" occur. But the operators, apparently believing that their instruments were wrong and that the reaction could be kept under control on January 21, did not shut the unit down. At other times, workers had managed to bring potential runaway reactions back into line, and they made the adjustments that had worked before. Their actions, which kept the unit running, had been condoned by the company over the years despite the dangers involved.

Between July 1996 and January 1997, at least four such events had been reported, and many operators told investigators that numerous temperature excursions had occurred where the unit was not depressurized. But when the reactor could not be controlled January 21, Tosco concluded that the operators were ultimately responsible.

What happened January 21
The reports presented to the Board of Supervisors paint a terrifying picture of the events leading up to the explosion, when a rapid increase in temperatures in the unit's Stage 2, number 3 reactor was masked by poorly designed instrumentation, malfunctioning equipment, false data, and conditions that changed far faster than anyone had previously seen.

Just the week before the blast, the unit had been shut down due to a tube leak on a heat exchanger. While starting up the unit January 17 after the repairs were completed, another leak occurred, this time on a clamp previously installed over a leak on a different reactor. Start-up procedures continued while the clamp was repaired. But on January 20 the reactor was shut down so the clamp, which had started leaking again, could be fixed. By then the plant was running, and the feed, consisting of hot oil and hydrogen, was simply increased to the other reactors to keep production going.

At the same time a new computerized temperature monitoring system for the reactors was taken out of service, after operators complained that it did not work. This equipment had only been in use for a few days, and company officials decided to reconnect the old system, known as a "data logger," while the unit was running so as not to interrupt the start up process. The company had known of problems with the data logger itself for weeks. Three of the five 1996 excursions reported were directly related to data logger failure. A second set of temperature indicators were located outside the control room, on a panel underneath the reactors.

Operators reported that the data logger began showing what they believed to be incorrect and confusing information about 10 minutes prior to the explosion. Temperature indications bounced back and forth, from low to high and back again; alarms indicating high temperatures sounded. The temperature controller, meanwhile, did not indicate abnormally high temperatures. The data logger was set to read "zero" when the top operating limit of 800 degrees was reached, but operators had never been informed of this fact. Because the data logger calculated average figures, these values of zero were averaged in. It appeared that temperatures had dropped, when in fact the excursion had spread inside the reactor. All this took place within 7.5 minutes.

Meanwhile, the temperature increase produced methane gas in the reactor, which mixed with the hydrogen gas normally in the system. A drop in hydrogen purity can be one indication of a temperature excursion. But the hydrogen analyzer, which would have shown a change in the composition of the gas, ran with a time lag of seven minutes, so up-to- date information that the operators could have used was not shown. When its alarm finally sounded, it was too late.

Between 7:34 and 7:38 p.m., operator Michael Glanzman, left the control room to check the temperature readings on the outside panel. His radio failed when he tried to call the control room from outside. Minutes later a pipe on the outlet of the Stage 2 number 3 reactor ruptured, and a huge explosion occurred. Glanzman was killed. The blast shook the refinery, blew out windows, and overturned office trailers in a wide area of the plant. While an enormous fire engulfed much of the hydrocracker, operators shut down the unit, depressured and secured it.

Rapid temperature increases have long been known to be a serious hazard of hydrocrackers, since heat is created by the process itself. Yet few such plants, including Tosco's 34-year-old unit at the Avon refinery, have automatic shutdowns that operate when critical temperatures are reached. Such a device, if it worked properly, would most likely have saved Glanzman's life and prevented the explosion. Automatic shutdowns are now being installed. New instrumentation has also been put in. The second set of temperature indicators has been moved from underneath the reactors to the control room. The unit is now up and running again.

Jim Altenberg is a member of OCAW Local 1-5 and an operator at the Tosco Avon refinery.  
 
 
Front page (for this issue) | Home | Text-version home