Based on historical data and updated with recent experiences, the system provides complementary inputs in the decision-making process, SS Bajaj informs R Srinivasan.
In India, as elsewhere in the world, deterministic models are used to assess events, the relative importance of features and safety measures in nuclear plants. SS Bajaj, Chairman, Atomic Energy Regulatory Board, spoke to us about the qualitative and quantitative measures that are adopted, accuracy of probabilistic approaches etc. Excerpts of the interview:
In terms of safety of nuclear plants, what are the qualitative and quantitative measures that are adopted here and abroad?
The safety objective of nuclear power plants is that these installations should not pose undue risk to public and environment. The risk due to nuclear installations shall be a fraction of the risk due to all other similar sources.
To meet this objective, a concept of defense in-depth is practiced for the design and operation of nuclear plants. The plants are designed robustly considering design principles such as redundancy, diversity, fail safe, single failure criterion, selection of proper materials, quality control in manufacturing, construction, commissioning and operation etc. This ensures safe operation of the plant at desired operating conditions. However, it is anticipated that plant conditions will deviate from its intended conditions during the life-time of the plant due to various disturbances. The plant has systems to bring back the operation to normal conditions. In cases where anticipated conditions threaten to become accidents, provisions exist in the plant to quickly bring it down to safe shut down conditions.
For those conditions which do cross into accident conditions threatening the plant's safety, enough engineered safety features are provided in terms of barriers and levels of defense in-depth to mitigate their consequences and bring the plant to an acceptable controlled state. Failures and events which could adversely affect the plant are postulated and design provisions are made to limit their progression and consequences. Emergency preparedness procedures are in place and well-rehearsed periodically to protect public in uncontrolled accidents, which by design have an extremely low probability of occurrence.
All the measures described above qualitatively have quantitative safety criteria and measures to assess their effectiveness. The primary quantitative measures require that doses to plant personnel and public due to all radiological pathways under normal operation of the plant are significantly below prescribed safe limits. Within these limits, a principle of ALARA is applied to keep the doses to 'As Low As Reasonably Achievable' under all circumstances.
At all other plant states acceptable dose limits are laid down on the basis that risk to the public is always very low in the zone of residual risk. Additional engineering criteria with sufficient safety margins are prescribed for each barrier and level of defense in-depth for design and operation of the plant so that the radioactivity in the core is kept well controlled, confined and in a cool geometry.
In probabilistic safety assessment (PSA), the failure rate of basic events based on historical data is taken into consideration as inputs. How accurate is this system? What are its advantages and disadvantages?
Probabilistic safety assessment (PSA) is a very useful tool to understand the safety of a plant, and is used as a complement to the more traditional deterministic approach. In PSA, the failure rate of basic events based on historical data is taken as ‘prior’ information, which is updated using the recent operating experience to reflect the current practices in operation and maintenance. This approach is used worldwide. As the operating experience increases, the data becomes more relevant and realistic to the plant. The system is effective in providing complementary inputs in the decision making, which is largely deterministic.
A combination of PSA and the Rasmussen report (US) helps in better safety decisions. What can we pick up from the Rasmussen Report?
The Rasmussen report was the first large-scale application of PSA (1975). After that, PSA methodologies have evolved and today most plants, including those in India, have their PSAs performed. So PSA studies complement the ‘deterministic’ approaches and provide an integrated insight into the safety of the plant, and help identify the safety importance of various features and aspects of design and operation.
In India, deterministic models are used whereas the worldwide norm is to use deterministic and PSA in combination. Should we do the same?
Worldwide, the traditional deterministic methods are predominantly used for licensing basis analyses. In recent years, PSAs are being used increasingly, to complement the deterministic approach. In India too, currently, regulatory decisions are made using the input from deterministic analyses and complementary insights from PSA in line with international trends.
PSA analysis helps better our understanding of safety perception i.e. there is Anticipated Transients Without Scram (ATWS) whereby a reactor, which should shut off automatically when a safety situation arises, fails to do so as in the case of Japan. To what extent are our plants addressing ATWS?
In the Japan accident, the reactors were automatically shut down following the seismic event. The damage occurred due to the subsequent failures in decay heat removal function. In India, for pressurised heavy water reactors, the design requirements is to provide two reliable, fast-acting, independent and redundant shutdown systems. Hence, ATWS scenario becomes ‘incredible’. For new reactors, apart from the active shutdown system, passive means are provided to insert negative reactivity in case the main shutdown system fails. Thus protection against ATWS is fully addressed.