**An example of e-PAR on web**

Fig. 11. Flow of functions and operations on e-PAR (browsing, editing and reanalysis)

### **3.4 CoolRep**

As indicated in Figure 12, at a top hierarchical level the key tools utilised are, in addition to Scarab for argumentation model development, CoolRep and a smart search engine.

'CoolRep' is an advanced, internet-based approach to management of documentation and providing an interface with users – both technical and non-technical. It allows the vast volumes of relevant information to be presented in a user-friendly manner, with different access options for different stakeholder groups. Since March 2010, the CoolRep 2010 version in Japanese language is available (http://kms1.jaea.go.jp/CoolRep/) and demonstrates the capability of JAEA (in the future, together with other Japanese R&D organisations) to support production and review of safety cases for deep geological disposal (Figure 12). An

A Challenge on Development of an Advanced Knowledge Management

assure fundamental applicability of this tool.

manuals (both conventional and video format).

González, 2006).

generation system.

**4. Discussion and conclusion** 

versions of information and data being used by different groups).

System (KMS) for Radioactive Waste Disposal: Moving from Theory to Practice 181

 During production, a single read-only master exists containing the accepted version; amendments of components may be produced in parallel, but contain digital signatures of the authors and are opened officially only after acceptance and digital signature of the report coordinator (assures implementation of the QMS and prevents different

So far, not all functions are fully implemented – but sufficient test cases are available to

As noted previously, transparent quality management is particularly important for the JAEA knowledge base. Even though guidelines are not yet agreed by all stakeholders, a demonstration of QA review for some text in CoolRep 2010 version and supporting kernels has been implemented. A further contribution to technical QA is the capability to solicit feedback from users. Such feedback can, in addition, be used to improve the structuring of information and the presentation software and play an important role in establishing dialogue with users, which includes the implementer, regulators, involved researchers, academics and the general public. CoolRep provides a portal that not only gives technical users access to KM tools and the KB, but also includes demonstration cases and user

In terms of the tools shown in Figure 12, progress has been slowest on smart search engine development. This is because an early decision to develop new software can be seen in retrospect to have been made without realisation of the extremely rapid rate of progress in this field. The effort is now focused on tailoring existing tools to the particular requirements of the JAEA KMS. In order to use smart search engines, it is critical to establish a clear vocabulary or taxonomy for the complex and multidisciplinary waste disposal field, in order to allow structured contextual searches that go beyond simple keywords. Much of the ongoing development in this area involves some form of semantic analysis, which has been claimed to be the key to 'next generation' knowledge management (e.g. Kawata et al., 2006; Umeki, 2007; Umeki et al., 2008). In order to form the basis to test such approaches, efforts have first concentrated on definition of clear ontology and examination of approaches to 'ontology cleaning' that have been used in other areas (e.g. Borrego-Díaz and Chávez-

The structured approach to initiating the development of a new KMS seems to paying off – tools and methodology have been seen to be applicable to a wide range of JAEA activities for deep geological disposal of radioactive waste and provide a better overview of context that was previously available. The need for a major change in approach of the JAEA activities to introduce KMS concept, processes and technologies/tools is now widely accepted by senior managers and younger staff, in particular, has provided positive feedback to the initiatives investigated. Nevertheless, in some critical areas – especially associated with the development of smart search engines – progress has been limited to defining project specifications and a key challenge will be assembling the support team to carry out the required software tailoring work. Probably most promising is the observation that, even at an early stage, the tools used provide hints of how increasing numbers of knowledge management functions can be automated, which will be a key to any next

English version is under development. Here, the name 'CoolRep' was chosen for this environment friendly option that minimises the use of hard copy, which was developed on the basis of lateral thinking by analogy to the CoolBiz campaign in Japan (http://en.wikipedia.org/wiki/Cool\_Biz\_campaign).

The CoolRep is produced entirely in electronic form and is provided on the internet in the form of a short, easily readable overview (2010 Japanese version is equivalent to around 75 pages) with extensive hyperlinks to:


Fig. 12. Links between CoolRep, the JAEA KMS and argumentation models

The advantages of this approach include:


English version is under development. Here, the name 'CoolRep' was chosen for this environment friendly option that minimises the use of hard copy, which was developed on the basis of lateral thinking by analogy to the CoolBiz campaign in Japan

The CoolRep is produced entirely in electronic form and is provided on the internet in the form of a short, easily readable overview (2010 Japanese version is equivalent to around 75

 support text providing more detailed technical input of major technical components (termed KERNELs – Knowledge Elements incorporating Requirements, Novelty,

both the generic, overall safety case structure and the component KERNELs in the form

further nested hyperlinks in both text and argumentation models that provide more

summary

Knowledge base -Database -Tech. reports -QA records - ・・・

Argumentation model

**Could corrosion rate increase with time leading to early failure?** 

**The long-term corrosion rates measured in experiments under relevant conditions are well below the reference values of 0.01mm/y**

> **Realistic corrosion rates from long-term corrosion experiments (less than 2μm/y).**

**The longevity can be specified by determination of an overpack thickness needed for mechanical integrity and adding an allowance for corrosion expected during the period for which integrity is required to be maintained.**

technical detail, including ultimately full text of key references

visual support material, including graphics, videos and animations

Knowledge management tools - Argumentation model editor (Scarab) - Expert system development interface - Integrated reporting system(PAIRS)

Fig. 12. Links between CoolRep, the JAEA KMS and argumentation models

responses, issue resolution forms) can be viewed by direct links

 The overview can focus on logical presentation of the safety case; technical support information is accessed directly when desired, rather than either clogging development of argumentation model with superficial detail or being isolated in abstruse specialist

Technical depth can be increased to the level desired by the user via the nested

 Wherever relevant, contained material is reviewed via a rigorous internal quality management system: comprehensive quality records (review comments, author

Knowledge supply Knowledge search

provide access to other knowledge management tools and knowledge base

(http://en.wikipedia.org/wiki/Cool\_Biz\_campaign).

pages) with extensive hyperlinks to:

Experience and Limitations)

of argumentation models

all review and QA records

user


Interaction


reports

hyperlinks

Knowledge access - documents - figures, tables, data - videos & animations - KM tools - QA records - ・・・

**CoolRep**

**JAEA KMS**

The advantages of this approach include:

provide access to relevant external websites.

KERNELs - PA

home



 During production, a single read-only master exists containing the accepted version; amendments of components may be produced in parallel, but contain digital signatures of the authors and are opened officially only after acceptance and digital signature of the report coordinator (assures implementation of the QMS and prevents different versions of information and data being used by different groups).

So far, not all functions are fully implemented – but sufficient test cases are available to assure fundamental applicability of this tool.

As noted previously, transparent quality management is particularly important for the JAEA knowledge base. Even though guidelines are not yet agreed by all stakeholders, a demonstration of QA review for some text in CoolRep 2010 version and supporting kernels has been implemented. A further contribution to technical QA is the capability to solicit feedback from users. Such feedback can, in addition, be used to improve the structuring of information and the presentation software and play an important role in establishing dialogue with users, which includes the implementer, regulators, involved researchers, academics and the general public. CoolRep provides a portal that not only gives technical users access to KM tools and the KB, but also includes demonstration cases and user manuals (both conventional and video format).

In terms of the tools shown in Figure 12, progress has been slowest on smart search engine development. This is because an early decision to develop new software can be seen in retrospect to have been made without realisation of the extremely rapid rate of progress in this field. The effort is now focused on tailoring existing tools to the particular requirements of the JAEA KMS. In order to use smart search engines, it is critical to establish a clear vocabulary or taxonomy for the complex and multidisciplinary waste disposal field, in order to allow structured contextual searches that go beyond simple keywords. Much of the ongoing development in this area involves some form of semantic analysis, which has been claimed to be the key to 'next generation' knowledge management (e.g. Kawata et al., 2006; Umeki, 2007; Umeki et al., 2008). In order to form the basis to test such approaches, efforts have first concentrated on definition of clear ontology and examination of approaches to 'ontology cleaning' that have been used in other areas (e.g. Borrego-Díaz and Chávez-González, 2006).
