The Prediction Method Of Fluctuat Operating Condition Parameters In Mechanical Equipment

The dissertation is based on the project provided by National Natural ScienceFundation,that is:Research on catastrophe failure prognosis of non-linear rotarymechanical system(50975105).With the development of technology, the equipment inelectric power and mechanic industry is getting more giant and sophisticated. The state ofsuch equipments plays a meaningful role in the safety of the enterprise and its economicproduction. The technology of prediction can dig out the changing rules of the equipmentfrom the development path of its historical statistics, so that potential breakdown cansomewhat be forecasted, which provides technological support for a properly-arrangedmaintenance plan and thus guarantee the safe running of the machine.For the volatility of mechanical equipment vibration data, a prediction model basedon grey theory and Markov theory is presented. Firstly, a new information GM(1,1) withconstant dimension is applied during the prediction of sample data. Then the Markov stateinterval is divided according to the error percentages between the measured state data andtheir grey forecast results, and the Markov state transition probability matrix is established.During the process of equipment state prediction, the Markov correction value iscalculated by the Markov state transition probability matrix and error percentage statevector of current state, thus, the prediction of volatility state data is realized by using theMarkov correction value to modify the grey forecasting results. Besides, in seeking fornew ways, this paper applies Ant Colony Algorithm to the prediction of potentialbreakdown, thus founding a prediction model of signal restructuring based on Ant ColonyAlgorithm for fluctuat operating condition parameters. With this model, vibration data isfirst used to vibrate again within certain range after the process of signal structuring. Thenthe idea of pheromone in Ant Colony Algorithm is introduced to predict. Last but not least,this paper presents the author’s research during his postgraduate study——the check andrepair information management system that the motor factory of a automobilemanufacturers, and explain how prediction of potential breakdown of the equipmentworks in this system.The example of the submersible pump proves that, the Grey-Markov prediction model for fluctuat operating condition parameters established in this paper, have goodresults in both the accuracy of the prediction and the tendency of the volatile data. Besides,the signal restructuring prediction model based on Ant Colony Algorithm for fluctuatoperating condition parameters, as a new prediction model, not only have a high accuracyof prediction, but also have good results on the prediction of the tendency of the data.More, through the prediction model established in the system, this method can be appliedto enterprises, in order to supply more information for the engineer to have a rightfuljudgment of the state of the equipment, so that the judgment is more scientific and morereliable.