Study Of Computerized Maintenance Management System Oriented Prognostics Approach

Due to technological advances and to increasing competitiveness of countries of low production costs, the industrial sectors of developed countries have to face constantly new challenges which are increasingly difficult. These challenges have as principal objective the maximization of competitiveness by the reduction of production costs, the augmentation of the installations profitability, and the creation of innovative products by guaranteeing staff and equipments security, and by respecting the regulations in terms of environmental requirements. The development of solutions capable of improving the production systems performances is then necessary in order to maintain the production sites survival at the heart of the developed countries. Industry is one of the engines of the economic development of a country.In fact, maintenance provides the possibility of exploiting enterprise resources in order to improve their performances by optimizing the utilization of human and material means. Since its beginning, maintenance has not ceased to progress and improve due to the emergence of Information and Communication Technologies(ICT) as well as due to the requirement and exigency imposed by the worldwide economic context. Maintenance has become a true discipline with its own methodologies and concepts. To make the classical strategies of maintenance more efficient and to take into account the evolving product state and environment, prognostic models need to be developed as a complement of existent maintenance strategies. When the maintenance strategy includes a prognostic function of the equipment remaining useful lifetime, we speak of Prognostics and Health Management(PHM), a domain from which has emerged the "PHM society".The prognostic is a quite new area of interest, it is the ability to ―predict and prevent‖ possible fault or system degradation before failures occur. Actually, if it is possible to predict the condition of machines and systems, maintenance actions can be taken ahead of time. As a result, minimum downtime can be achieved. Prognosis has been defined as ―prediction of when a failure may occur‖ i.e. a means to calculate the Remaining Useful Lifetime(RUL) of an asset. In order to make a good and reliable prognosis it must have a good and reliable diagnosis.Considering neural network(NN) based remaining useful life(RUL) prediction approach. A new performance degradation index is designed using multi-feature fusion techniques to represent deterioration severities of facilities. Based on this indicator, back propagation neural networks are trained for RUL prediction, and average of the network‘s outputs is considered as the final RUL in order to overcome prediction errors caused by random initiations of NNs. Accurate equipment RUL prediction is the key to effective implementation of condition-based maintenance(CBM), which aims to prevent unexpected failures and minimizing overall maintenance costs. RUL prediction has been implemented using either a model-based or a datadriven method. Data-driven methods, requiring no special product knowledge, are based on the fact that condition monitoring data and the exacted features vary with the development of initiation and propagation process, or the degradation process.In this thesis, a systematic data-driven prognostics method based on ANNs was presented. Feature extraction was performed by time-frequency analysis. Typical features were selected using competitive neural network. A new degradation indicator was derived from neural networks after typical features being extracted. RUL was predicted by constructing a BP neural network based model.. Finally, an experiment is set up based on a Bently-RK4 rotor unbalance test bed to validate the neural network based life prediction models, experimental results illustrate the effectiveness of the methodology.