Determination of the Probability of a Gear Pump Fault Using the Bayes Network and Netica Software

Abstract

The literature dealing with risk analysis and finding an adequate solution in the mechanical engineering sector shows that the use of the Bayesian network as one of the methods for risk assessment in this context is well known. Therefore, the authors of this paper use the Bayesian network, while accordingly, for the purpose of simpler work with Bayesian networks, special software has been developed, one of them being Netica, as the most commonly used method for determining an adequate solution, and in this case determining the probability of malfunction. gear pumps if the probabilities of variables that directly affect the correctness of the pump are known. The procedure for determining the probability of malfunction of the gear pump in the specific example shown in this paper is obtained first by designing the Bayesian network, and then by creating the same network in Netica software, where pre-defined nodes (variables) to which assigned values are entered in program, after which the arrows show the dependences of the given variables, and then enter the conditional probabilities of these variables and form a network on the basis of which the probability of malfunction of the gear pump is determined.