luni, 29 octombrie 2018

DEVELOPING AND USING NON-PERFUNCTORY KNOWLEDGE-BASED SAFETY TEMPLATES-TAKE AWAY SAFETY EXPERTISE FOR SME


There are many economic units (in Romania at least) that are not into routine production. These units need safety also- and the present paper describes a possible solution for them. Templates may seem trivial in the experience of many developed countries- but they are not simply templates but expert system-based ones.

Templates are- in our vision:

-instruments to check in specific safety situations;

-instruments for learning;

-best safety procedures in order to do a specific task without injuries;

One take-away template is composed by:

-A Well Balanced Safety Template Input;

-The expert mechanisms;

-The dissemination of results dashboard (DDR)

So, on the input of the template there are work facts derived from the necessities of the worker. Why is this template input interesting? We call it Knowledge Based Template Input (KBTI) and we are considering as an input for the template also not measurable facts – facts that were identified but on which the worker could express – with a degree of incertitude- of course. We are speaking about Well Balanced Template Input when the coefficient of uncertainty is >=0.8 – the ideal value being 1. The template auto-computes these uncertainty indices and could give some advice to the filler employee- like ~you need more data~ or ~please focus on the estimated pressure of reservoir 1 when the event is happening~. Moreover, the template records the input data as it is transforming it into a pattern for lessons well learned.

What happens inside? Inside the expert systems offers solutions arranged as follows:

-For immediate use- the solution assures the safety and health of the employees but is not guaranteeing the integrity of the machines or processes or the environment; templates for immediate use are activated by an expert system module with the capacity to learn and analyse the immediate decisions made using these templates;

-For quick implementation-the solution assures the safety of the workers and the integrity of machines and logistic. It could assure also the safety of the work environment, if the environment has not a certain degree of complexity;

-For reflection and interpretation- the solution protects all the three elements of the workplace but is costlier and sometimes difficult to apply. This solution is using a prospective expert system that would analyse also the possible uncertainty based solutions;

Employees are assuring after everyday use the necessary feedback like:

-Template is fit for our work problems- enounce 3 major advantages;

-Under development- we could use part…;

-Not comprehensive;

-Not understandable;

-Wrong,

Why we want to present such templates at Hazards XXXIX? Because more and more companies would became”non-traditional” ones- and would need to use such an expertise and to base their safety to expert decisions and not just on static knowledge.

MASTERING LITERACY OF SAFETY USE CASES- HOW TO TAME THEM, HOW TO MAKE THEM EFFICIENT, HOW TO RE-VALUE THEM


A safety use case describes in Romania the development and consequences of an unpredicted work event, a near-miss situation and things like this. Traditionally, use cases are having an educational purpose. Safety use cases are narrations of real unexpected events- that were recorded using the recollection of survivors together with the expertise of safety inspectors. Sometimes, even the use case is very interesting; it is not recorded because it is considered as not significant.

Our approach is oriented towards the improvement of the development process of safety cases- not just around safety training but also focusing on the optimal decisions that should be taken by the top management connected with the safety and well being of the workers.

We have started with the improvement of classic use cases – UML style. Here beyond the traditional components we have introduced two distinct classes- ”Safety affected”- that is leading to the unexpected event- and could be further developed using the 5 Why method – and ”Safety mitigated” including mitigation solutions.

Our further developments have established a cognitive framework for the safety use cases- with three main parts:

-the description (in natural terms) of the unexpected event start-up- having attached the results of the 5 Why analysis and also a taxonomy that is revealing the exact role of the Human Failure, Safety Failure and Machine Failure in the birth of the unexpected event.

-the happening of the unexpected event;

-the resulting consequences and the possible scenarios that could happen or that could be developed in the making of an optimal decision;

The safety use case framework is boosted by two expert mechanisms- an use case constructor- that is performing the most  routine operations regarding the optimized building of the use case (the use case constructor-UCC- could be goal oriented considering simultaneously 3 different goals) and an use case destructor-UCD- that is ”killing” the non-interesting or non-significant parts. 

UCC is completing/tailoring the terms from the narration with terms found in a specific ontology developed for this instrument.

A special accent is put on the so called “transition areas” when the use case is performing its main developments towards the nefarious event.

Having a draft use case developed by UCC and not destroyed by UCD- the case is tested against uncertainty factors that are specific to the industrial domain of development. An old use-case could be”refurbished” and re-used.

 The obtained results have been integrated into a primary list of safety use cases. The use cases will not cover all existing problems but will focus on the most important ones which most likely will have a big impact on the decision making process.

We considered interesting to share this experience with other interested parties at the Hazards conference- hoping eventually to establish a cooperation regarding the development and usage of safety use cases.