The Guangdong Daya Bay nuclear power station (GNPS), which houses two Framatome 900 MWe PWR units, came into commercial operation in 1994. Since then, GNPS has continuously pursued the improvement of fire safety. One of our ambitions has been to set up a performance measuring methodology enabling a more quantitative dynamic evaluation of plant fire safety to be made.

The result is a set of six fire protection performance indicators (FPPIs) which provide a yardstick to check all the essential levels of the plant’s defence-in-depth philosophy and the direction of performance trends.

The FPPI’s are described in detail in the panel below and overleaf.

FPPI ANALYSIS AND REPORTS

Except for indicator No 1, Fire, for which no event has been recorded up to now at GNPS, all the other indicators have provided useful databases for both root cause analysis and performance trend analysis.

For any particular incident, safety engineers must conduct an analysis in accordance with the relevant procedure. From the overall operation event management point of view, each event can be categorised as a licensing operation event (LOE) or internal operation event (IOE). In either case, efforts must be made to identify the root causes. These may include failures in engineering aspects, such as design, manufacture, erection, or managerial aspects, such as training, procedures and work habits. The trend analysis focuses not only on comparing the indicators over different periods of time, but also on the distribution patterns of the causes, the potentially vulnerable areas and the shift of responsibilities. As a result, we can take corrective actions and set up improvement programmes in a reasonable time frame and with an appropriate allocation of resources.

For all six indicators a monthly synthesis report is issued inside the company. This report is accessible to the regulators. In addition, a presentation is made at the monthly meeting of the station’s direction team, attended by all the station managers and branch heads. Every three months, an evaluation report on plant fire safety is issued and presented to the Plant Nuclear Safety Committee (PNSC). Finally, an annual report is produced summing up the entire year’s performance and indicating the potential areas for improvement for the coming year and, if necessary, for longer-term programmes.

The FPPIs and relevant data are accessible through the company’s computerised Safety Information System Network.

LEGAL BASIS and FPPI concerns

China’s nuclear safety legislation has no specific legal definitions on fire incidents, nor as far as we know, is there a standard uniform reference in the international nuclear industry, so our FPPIs have to make reference to existing domestic legislation.

China’s Public Security Ministry wrote the current official fire incident definition and classifications; for the power industry, China’s former Electric Power Ministry produced definitions of fire in power plants. However, these definitions and classifications are of conventional fire concepts.

With reference to these legal requirements, the FPPIs take into account the special concerns of nuclear power stations. These concerns include the impact of fire on nuclear safety and its relationship with the public, particularly the local communities.

Regarding nuclear safety, fire is an unpredictable threat, which depends more on which safety function fails than simply the scale of the fire. As for public concern, we must take account of the fact that the report of an insignificant fire may cause misunderstanding or overreaction. The keynote we set up for our programme is prevention and, like the proverbial iceberg, we want to discover what lies below the water surface.

FURTHER IMPROVEMENT

The FPPIs have already proven a useful tool, revealing areas for further study and improvement.

One of the critical issues is to understand the position of a fire in the cause-and-effect chain, or what kind of role a fire plays in the sequence of an accident. For example, poorly carried out hot work may cause a fire that can be regarded as an initiating event in the sequence of an accident. Fire of an electrical appliance may be only the consequence of a poor design or of a maintenance error on that piece of equipment, but it may become a magnifier of that consequence, leading to a more severe and even uncontrollable situation. In such a case the initiating event is that electrical appliance failure instead of the fire.

The point in question is how FPPI can be further developed to distinguish fire which is no more than a conventional incident from the fire which has realistic potential threat to common mode failure. If this information is available, we are certainly to have a more in-depth, more realistic assessment of operations fire risk and thus more effective fire protection programme.