Study On The Method Of Tendency Forecast For Marine Gas Turbine

Gas turbine is an important part of the ship’s power system. In order to protectthe ship’s safe navigation and perform tasks, we should monitoring and predict thestate of the gas turbine. According to the results predicted, we can repair the gasturbine in advance to control fault at an early stage. It do benefit to reducemaintenance costs and maintenance costs. This paper mainly predicts stateparameter of gas turbine to achieve life prediction and forecasting task executioncapabilities.This paper studies several commonly used predictive algorithms, and test andverify them. Its basic theory and the derivation of the formula in-depth analysis,and thus prepared the program. The paper also analyzes the advantages anddisadvantages of these predictive algorithms.Secondly, in order to predict with greater accuracy, we use a combination ofpreferred and virtual predictable ideas, which are used to predict the state of gasturbine. So the paper verifies the operability and correctness of the combination ofpreferred and virtual predict, it is superior to direct prediction.The paper also sumup gas turbine failure mechanism. And this article describes some datapreprocessing methods, and verify them.Finally, Based on the entire life-cycle gas turbine degradation data that appears,come to the actual gas turbine will be running pre-and post different degradationtrends., so the paper uses S-W test and skewness test to find the transition point.And these methods are feasible. Then this paper proposed normal and t distributioninterval prediction, and compared them to find normal distribution intervalprediction is good. Based on this, declining trend on the compressor efficiency ismade predictions, to determine the left life of the compressor. In addition, the paperaccording to the actual needs made a preliminary analysis about gas turbinecapacity forecasts base on "recession rate and speed related"...