Prognosis and Detection of Experimental Failures in Open Field Diesel Engines Applying Wienerˈs Artificial Immunological Systems

Abstract

The high costs of open-field diesel engines arise from the lack of maintenance of these systems. Thus, the maintenance of this equipment has been treated as a great challenge, as some methods of data monitoring are not possible to be implemented, given the inadequate sensing conditions, plant location, local climate, facilities, even the methods and maintenance routines. In a second step, the labor is not qualified and of sufficient quantity to meet the demand, resulting in a slow and inefficient system. One of the challenges of predictive systems is to inform damage and failures in real time of the operating conditions of these machines and equipment. This work demonstrates the possibility of analyzing and detecting failures in open field predictive systems, using the concepts of vibration and acoustics in artificial intelligence. One of the results of this work demonstrates the robustness of the negative selection artificial immune system algorithm, whose application of the Wiener filter was of fundamental need. The other result demonstrates the versatility of conditioned use both or just one of the concepts between vibration and acoustics, in prognosis and fault detection. Considering the versatility of using these two techniques, it is possible to affirm that, the predictive systems of real time analysis have an effective solution directed to the area and, if implemented, it is of low cost and high efficiency.