The Process Can Be Managed

Human ingenuity can now create large-scale technological systems whose accidents rival in their effects the greatest of natural disasters, sometimes with even higher death tolls and greater environmental damage.

Exhibit A: Thursday’s serious accident at a nuclear processing plant in Tokaimura, Japan, that sickened dozens of workers and area residents and caused some local evacuations. Because of radiation leaks--levels near the plant were reported to be 15,000 times normal--thousands of nearby residents were warned to stay indoors.

According to preliminary reports, the accident was caused when workers mixed “too much”--nearly eight times the normal amount--of uranium in a storage tank. This “human error” triggered a “critical-mass accident”--an uncontrollable chain reaction that unleashed an intense burst of radioactivity. This wasn’t the Tokaimura plant’s first major accident; a “very serious” fire and explosion occurred on March 11, 1997, injuring 37 workers and releasing radiation.

Two weeks ago, I was in Japan to attend an international conference on human-factors research in nuclear power operations. An analysis was presented of the human errors that occurred in Japanese nuclear power plants from 1966 through 1995. Among 863 incidents and failures reported to the Ministry of International Trade and Industry, 199 events at 49 nuclear facilities were blamed on human error.

However, as I pointed out, attributing accidents to actions of front-line operators is an oversimplification that only results in blaming the victims. One could reasonably argue that the reason behind the two major accidents in the span of 2 1/2 years at Tokaimura is the safety culture of Sumitomo Metal Mining Co., the owner and operator of the plant.

The safety culture is defined as the assembly of characteristics and attitudes in organizations and individuals that establishes safety as an overriding priority. Three general components of the safety culture are the appropriate organizational structure, the responsibility of top management and the attitude of staff at all levels. The requirements of individual employees are a questioning attitude, a rigorous and prudent approach and communication.

A common characteristic of complex, large-scale technological systems of nuclear power and processing plants, chemical plants and refineries is that large amounts of potentially hazardous, flammable, combustible or pressurized materials are concentrated and processed in single sites under the centralized control of a few operators. Both the performance and the inherent accident potential of such plants are functions of the way their parts--engineered and human--fit together and interact.

My research has shown that on many occasions, errors and resultant failures are because of a multitude of factors such as poor workstation and workplace designs, complicated operational processes, unbalanced mental and/or physical workloads and inadequate staffing, unsafe working conditions, faulty maintenance, ineffective training, lack of motivation and experience, nonresponsive managers, poor planning, rigid job-based pay systems and haphazard response systems.

Although accidents are always tragic and can have devastating effects, there are many lessons that society and industry can learn from them. As American philosopher William James has said, “Great emergencies and crises show us how much greater our vital resources are than we had supposed.” We can and should be able to utilize our vital resources to operate nuclear power plants safely; we do not have a choice in the short run. Yet we also need an overall paradigm shift in dealing with nuclear plant design, construction, operation and regulatory oversight.

To improve the safety problems of nuclear power, we need to improve the safety culture of this industry and address human and organization-related factors. To keep the nuclear genie around the world bottled up, we need genuine international cooperation and coordination among equipment manufacturers, operating companies, operators’ unions, regulatory agencies, international organizations, scientific and research communities and governments.

Najmedin Meshkati, an associate professor of civil/environmental and industrial and systems engineering at USC, conducts research on the safety of large-scale technological systems. E-mail: meshkati@usc.edu.