Using an expert system for the solution of operational failures in sludge pumps in oil well drilling equipment

Abstract

It has been developed and we proposed to implement an expert system that acts as a permanent technical assistant that contains a knowledge base, a bank of solutions and a history of failures, which allows to solve the failures that occur in the sludge pumps, when the machinery, that form the drilling equipment, is working. We estimated that the system can reduce the downtime, caused by failure, using the advantages offered by this branch of artificial intelligence that simulates the process of learning, memorization, reasoning, communication and action to solve problems; binary trees being the main data structure that was implemented. We estimate that the training time, for the handling of the expert system, is near to eight hours for personnel related to sludge pumps. In particular, the solution of a fault such as the loss of communication with the J box, the expert system would be solved in approximately 60min, while in the traditional way it could take up to four hours, depending on the availability of the specialist in charge of the drilling area. Therefore, the saving in solution time, restart and, therefore, economic resources is highly significant.