Материал: Крючков Фундаменталс оф Нуцлеар Материалс Пхысицал Протецтион 2011

Article XII of the IAEA Statute defines the types of activities undertaken to implement the safeguards. According to this article, the Agency has the right:

∙to verify the information on the design of nuclear facilities;

∙to demand that relevant national authorities keep records as required for NM accounting and provide access to such documents by the Agency;

∙to demand submission of the accounting documents to the Agency;

∙to send IAEA experts to the states that recognize the safeguards.

As mentioned already, for the purposes of the safeguards to be attained, the IAEA checks on the nuclear materials during inspections. To this end, it is necessary to know, first of all, which and how much material should be checked to allow a conclusion on the absence of NM diversion. Such verification involves three reference parameters:

∙significant quantity;

∙time of detection;

∙probability of detection.

Significant quantity of nuclear material

In the context of a national system of safeguards developed mainly to protect NM from theft or acts of terrorism, the word “significant” may be applied to a relatively small quantity of material, considering, for example, its high toxicity (plutonium). However, the international safeguards are intended primarily to make sure that governments do not procure nuclear weapons or other nuclear explosive devices (NED), and it takes a relatively large NM quantity to produce even one NED.

In practice, the safeguards apply to NM with various Pu and U concentrations and of various isotopic compositions. This fact suggests the need for defining the “Significant Quantity” (SQ) f or each material category, so as to indicate how much material should be looked for to determine whether there was a diversion in violation of the safeguards.

In general, the SQ is defined as an approximate quantity of nuclear material large enough to allow producing a nuclear explosive device, whichever conversion process is used.

The SQ varies depending on whether this material may be directly used for NED production, or it needs to be converted first, e.g., by enrichment (in which case it will be referred to as an indirectly usable material).

For directly usable materials, the Significant Quantities are established by experts of the Standing Consultative Group of the IAEA for Implementation of the Safeguards as follows:

66

∙8 kg of Pu (containing < 80 % Pu–238);

∙8 kg of U–233;

∙25 kg of U–235 contained in HEU.

For indirectly usable materials, the Significant Quantities are:

∙75 kg of U–235 contained in LEU (10 t of natural ur anium);

∙20 t for depleted U and Th.

Significant Quantities in inspection practice

The actual purpose of inspections at a specific facility does not necessarily lie in detecting diversion of one SQ (nor is it practicable in some cases). For instance, checking on nuclear materials at reactors consists primarily of counting, identification and measurement of individual accounting items, i.e. fuel assemblies. Once such items are counted, the purpose of inspectors is to find whether at least one FA is missing. Before research reactors of the IRT type were converted to lowenriched uranium (LEU), fuel assemblies contained 300-400 g of highly enriched uranium (HEU). It means that the inspection target in counting such items is smaller that one SQ (25 kg) for HEU.

At large facilities where nuclear material is found in bulk, it may be the other way around. In such cases, inspection will necessitate measuring large quantities of NM found in different physical forms and varying in chemical composition, including such material as waste. It may be tentatively assumed that the NM measurement error is about 1 % of the total measured quantity. It is clear, however, that 1 % of the inventory or throughput of a large facility with NM in bulk may exceed (by far in some cases) 1 SQ.

Let us consider a case where the IAEA safeguards cover several fuel fabrication facilities, with a large quantity of indirectly usable bulk material (LEU) being processed. Measurement errors and other factors make it necessary to set the inspection target at five SQs. In consequence, at such facilities there is no ruling out the possibility (with a credibility level of 90– 95 %) of one SQ diversion in the case of LEU. It does not mean, however, that diversion of one SQ is not detectable at all. Detection is possible even in this case, but its probability is lower. Besides, a high threshold of detection is offset by using additional means of control, such as containment and surveillance. Finally, it should be borne in mind that diverted LEU may be discovered at other stages of its conversion into a weapons-grade material.

67

In countries with a well-developed fuel cycle, it is theoretically possible to divert a significant quantity of material by adding up under-SQ material quantities diverted at each one of many facilities. With such an approach, diversion would involve materials of different types and categories and would necessitate concerted actions of personnel from a large number of facilities. That is why the IAEA regards this scenario as unattractive for a rogue state and highly risky. Besides, an attempt to counteract such hypothetical diversion scenarios would necessitate an inspection regime much too cumbersome for implementation by either a state or the IAEA.

Early detection

Early detection is a notion of importance to the safeguards. The earlydetection requirement determines the frequency of inspections at facilities. The system of the IAEA safeguards is based on the assumption that diversion of a significant quantity of nuclear material should be detected at an early stage, irrespective of whether it is committed once or on multiple occasions. The practical implications of early detection depend on the type of the nuclear material covered by the safeguards. In order to determine ‘timely detection’ in quantitative terms it is nece ssary to have a notion of the ‘conversion time’. The time of NM conversion is the time required in optimal conditions for converting NM in a given form into a component of a nuclear explosive device. The following typical conversion time periods were established on recommendation of the Consultative Group:

Besides, following the CG recommendations, the IAEA established the ‘detection time’ to be of the same order of magnitu de as that of the conversion time. The ‘order of magnitude’ here is a factor equal to 1–3.

In practice, it is sometimes difficult to stay within a short time of detection. So, at some facilities where nuclear material is found in bulk it is a far from easy business to reconcile a short detection time with the requirements of normal operation. For plutonium and highly enriched uranium (whose conversion takes 7–10 days), inspect ion targets are set at the upper boundary of the permissible time range (3 weeks). In those cases where acceptable inspection frequency falls short of ensuring early

68

detection, additional measures are taken to attain the required detectability, such as application of seals and surveillance.

Detection probability

Considering the global scale of the IAEA activities, it is an extremely difficult task to discover NM diversion with a 100 % probability and as early as required. That is why the Agency seeks to build such a system of safeguards that will meet these objectives with a certain probability. This probability should also be specified. Neither INFCIRC/66, nor INFCIRC/153 refers specifically to the level of confidence in detection, but the Agency interprets these documents as if this notion were covered there. From the viewpoint of the Agency, the detection probability should be so high as to keep a state from resorting to diversion and to ensure the

required confidence of the international community. The target detection probability was set by the IAEA at the level of 0.9÷0.95.

Extent of independent inspections

The effect of the probability of detecting a significant quantity of NM on the extent of IAEA inspection activities can be illustrated by a model case.

Suppose we have as many one-type articles as N = 12000. The SQ of the material under consideration (e.g., plutonium) is М = 8 kg. The mass of each article is m = 80 g. It is necessary to determine the number of articles to be verified during an IAEA inspection, if the probability β of attaining the safeguards purpose should not be less than 0.9.

The number of diverted components x in sample n is taken to be a random quantity whose distribution follows the hypergeometric law. The sample size is then determined by the formula:

Condition (3.2) implies that relatively large components are under consideration. Once fulfillment of this condition is checked upon, formula (3.1) will be used to find the number of measurements to be made during the inspection:

69

n = 12000 × (1 – 0,1 1/100) = 276.

Apparently, the extent of sample measurements is vastly reduced in comparison to the initial scope, if the adopted SQ is largely in excess of the component mass, and the probability of attaining the purpose is noticeably smaller than 1.

Equipment involved in implementation of the safeguards by the IAEA

The complexity and diversity of NM-containing facilities call for use of various techniques and equipment [4]. Table 3.1 presents the main verification procedures carried out by the IAEA at such facilities.

Since the 1980s, inspection activities have tended to involve expanded means of containment and surveillance, as well as non-destructive test methods.

70