Decision support system applied to forging processes

Long considered as a process for obtaining raw parts finished by machining, forging today enables the production of high value-added parts. For example, the manufacture of net-shape parts (forge-finished parts, with no need for finishing), the improvement of mechanical characteristics in service (fatigue resistance, etc.) or the reduction of mass and/or volume, all other things being equal, are all areas of development for this plastic deformation shaping process. These considerations have become possible thanks to the improved quality of production resources and their control systems. The implementation of increasingly technical product ranges requires recourse to experts in production, design or methods.

Paradoxically, attachment to a company is less and less frequent, and management turnover is increasingly rapid. As a result, knowledge and know-how linked to individuals diminish within the company each time an expert leaves. Training new referents takes time. Knowledge is generally capitalized on by project or following best practice advice, but this is not always adapted to the company's real needs in terms of responsiveness and productivity.

Most of the time, decisions in forges are taken by experts on the basis of their experience and analyses of similar cases, with trial and error to correct unforeseeable errors. Seeking information from experts whose availability is limited, or from company archives, is often time-consuming; hence the idea of using artificial intelligence to develop forging decision-support tools.

Based on an expert system, these tools consist of a knowledge base on the one hand, and an operating mechanism on the other. Search engines speed up access to information to help make the right decisions. The knowledge required for this can be capitalized on by experts upstream, and as and when it is acquired. Expert knowledge and its exploitation can then be dissociated.

The article begins with a presentation of the overall context and the basic notions of the approach; forging and decision support systems are illustrated. The beginnings of decision support systems applied to forging are discussed, before turning to two more recent examples of development. The first deals with the creation of routings based on business knowledge, the production resources available in the forging shop and their capability. The help provided facilitates feasibility studies and enables quotations to be drawn up fairly quickly. The second deals with mastering manufacturing processes. The assistance provided helps to identify the parameters to be monitored to detect drifts in real time, and to correct them to avoid scrapping parts.