ANALYSIS OF AI POSSIBILITIES FOR ASSESSING EXPLOSION-PROTECTED EQUIPMENT

The paper presents aspects of analyzing the AI possibilities in evaluating equipment for explosive atmospheres. In the first part of the paper, the context of the use of equipment in potentially explosive atmospheres is presented. The presence of flammable substances in industrial applications leads to additional requirements for technical equipment used in such spaces. Furthermore, machinery used in such spaces must remain in working order even when the risk of explosion is present. In summary, to confirm eligibility, equipment used in such spaces must be assessed, tested, and certified to ensure it functions safely about the risk of explosion [1] and [2]. The second part sets out the testing framework considered. On this occasion, some representative tests encountered in the assessment process, applied to AI agents for the purpose of identifying the assessment performance of equipment intended [3] for use in explosive atmospheres, are presented. The types of tests considered are those in which the following is assessed: understanding a set of requirements in the standard, and assessment of the fulfillment of a set of requirements in the standard. The forms in which the information is retrieved are plain text, a picture with a calculation formula, and a picture with a drawing. The options considered are binary alternatives and multiple alternatives. In the third part, the results of the tests are presented and discussed. An important conclusion of the approach is that each AI agent exhibits an individual score in terms of performance in evaluating equipment with the type of intrinsic security protection and the ability to convey the correct information.