Desarrollo de una herramienta de análisis de riesgos en operaciones industriales

Résumé

Any productive process is exposed to many risks that difficult to obtain the product according to desired specifications. It does not matter how monitored the system is, uncertainty is always present as a consequence of errors in calculus accuracy, wrong equipment calibrations, equipment limitations, or human errors even. Obviously, risk can’t be avoided on a whole way. Clients establish an error range, i.e. size or weight, which must be greater than production errors derived from the problems talked befote. We can assume the risk of having errors always this does not mean a reject of client delivery, with costs and loss of company image it would suppose. In the iron&steel industry case, the smaller the client tolerance is, the greater uncertainty becomes taking into account that production planning is changing and flexible time a time, avoiding big series and predefined schemas. This problem is not specific in one factory or in one property monitoring (hardness, thickness, etc.) but is generic and involves the full aspects of the process. In these circumstances, technicians must decide continuously if a product can be sent to the client in spite of a test can indicate a big error, or must be removed meaning a great entailing. In this work is presented a system development to value the risks derived from not fulfil client requisites, applicable to the whole processes and product features generically. The system is based on the next features: 1. It must be applicable to any kind of problem. 2. The characteristics of the variables introduced in the problem are not known, so experience-based techniques are not suitable. 3. Because of the same reason, it is no possible to introduce expert knowledge directly. 4. Applying the tool must not require great technical knowledge. This system was developed for a well known company of this sector, conceived as a decision maker helping system, through a risk management system in industrial processes.