Se afișează postările cu eticheta ontology. Afișați toate postările
Se afișează postările cu eticheta ontology. Afișați toate postările

duminică, 16 august 2015

SAFEONTO- Using ontologies to improve safety learning



Abstract

The paper follows the development of a general purpose safety ontology (SAFEONTO) that could be used as a tool for safety learning and also a framework for the study of various accidents and incidents happened at work- turning this unexpected events into lessons learned.


Development of SAFEONTO


Based on previous experiences[i] we have tried to develop a general framework for risk and safety, framework that could be expanded as wished by any safety expert. The idea was to have a referential that could be generally used in order to provide an instrument for training and analysis for specific workplaces.
In order to do this we have used knowledge maps built using CMAP TOOLS- and also the PROTEGEE tool for the ontology development.
What should include a general knowledge map?
                -details about the economic activity that generates risks;
                -data about the worker(s) and the machines used in order to perform that activity;
                -some details about the workplace and its specificity;
As we have more data- a transition from general to specific would be made easy. Figure 1 shows the concept behind the ontology. All these building elements are used to develop best-worst case models- that allow us to establish the optimal safety requirements.

Figure 1 Main concepts of the ontology

Figure 2 shows the knowledge map developed as a start-up for the ontology.


Figure 2 Start-up knowledge map

Any economic activity (that could be specifically described) is done by workers (employees) using tools at the workplace. Such an activity is generating risks.
On the next step we have introduced the consequences of the risks, as shown in Figure 3.


Figure 3 Risk development and consequences

One step further we have introduced the safety lane.



Figure 4. The activity and safety lanes

Now we have a functional framework for the ontology. This framework could be developed further as far as is necessary.
On the last step of development we introduce also a use case lane.






Figure 5. The complete knowledge map

The use case is taken from an investigation report and could be automatically transformed into lesson learned.
A further step in the development process would be turning the ontology into a decision assistant system.
The next figure shows the list of concepts.





Figure 6.List of concepts introduced in the ontology



CONCLUSIONS

 The development of ontology was an interactive process. We have used the experience of five of the best specialists in safety inside Romania in order to be able to capture the most important parts – the one that are allowing the generalisation used inside the ontology.
Our instrument could be used:
1. to analyse specific unexpected events that are occurring at the workplace;
2. to transform them into lessons learned- in order to improve safety and health at the workplace;
3. to develop specific models that could be used for the decision process;
4. to use the instrument for safety training;
As the main benefits:
-the ontology is model-based, allowing integration of design and safety analysis;
-safety considerations could be introduced early in the development phase
-use of Semantic Web and Ontology standards;
-semantics based interoperability at a language and tool level;
-Web-ready;



[i] http://safetyinknowledge.blogspot.ro/2014/10/safety-eduwork-and-risk-ontologies.html

marți, 10 martie 2015

RIONT-A MODERN RISK BASED ONTOLOGY


We have begun to develop an occupational risk based ontology- using as a main instrument for development Stanford University Protegee.
As soon as we are going to find a free host our ontology would be accessible for free on this support.
The specialists that are willing to cooperate in the development of this ontology- are gladly expected.


DEFINITIONS

We have considered some main start-up definitions as main pillars in the development of the ontology.
1. HAZARD- A hazard is any source of potential damage, harm or adverse health effects on something or someone under certain conditions at work.
Basically, a hazard can cause harm or adverse effects (to individuals as health effects or to organizations as property or equipment losses).
2.RISK – is the umbrella under which all occupational risks can exist.
3.INHERENT/INCIDENTAL RISKS – contain all risks that can either exist or be created by controllable actions (incidental), or exist due to uncontrollable circumstances (inherent)
4.EXTERNAL/OPERATIONAL RISKS – external risks can only represent external financial risks to an enterprise. External risks can be members of all other risks and must be a part of any discoverable risk set. Operational risks can only represent internal financial risks to an enterprise. Operational risks can be members of all other risks and must be a part of any discoverable risk set.
5.RISK FACTORS- Factors that  describe specific risks. Other definitions are following. We will try to use EOSH and NIOSH definitions.


STEPS IN DEVELOPING THE ONTOLOGY

We have started with three main classes:
1. Risk;
2. Risk factors;
3. Personal factors that are affecting risk;
We have added subclasses to these main classes.
We have defined instances where we were thinking that such instances are appropriate.


THE DEVELOPMENT PROCESS

The development process is actually underway. 
Some images from the development process are presented below.

Figure 1. The class browser of Protegee

Figure 2. Sub-class definition 


Figure 3. Vertical arborescent image

Figure 4. Spring image 

Figure 5. Horizontal arborescent image


CONCLUSIONS

At this moment we were not able to find something very useful by developing such an ontology- as it has very little space to document the terms, it is not accepting links and has a somehow rigid format. In this respect we are going to link the ontology with knowledge maps in order to obtain a better learning instrument. 
Some images from the running of the ontology are presented below.
Figure 6. Main screen


Figure 7. Risk factors


On a larger scale it is interesting that such an ontology could be exported as CLIPS code or as a Java code, being able to be embedded in such programs.