Reliability Design Optimization Of Train Gears Based On Approximate Model

Gear as a drive mechanism has been widely used in various transmission systems. Itdepends on teeth profiles contaction for transmitting motion and power betweenrandom angles of shafts, and it has advantages of transferring power, high efficiency,accurate transmission ration, longevity service, and working reliable. Now, the speedof train has been vastly improved in China, and as drive train parts, the transmissiongear will work at the circumstance of heavy load. So the tooth surface veneer、pittingproblems have come out in some trains. Based on the fact, analyzing of stress of gearsand optimization of mesh parameters will be imperative.At the first, the parameterized model of train gears was established by the finiteelement software ANSYS parameterized design language APDL, and then analyzed thetooth contact stress and tooth root bending stress under the equivalent maximum power.At last, under the restrains of reliability, took the method of non-dominate sortinggenetic algorithm (NSGA-II) which is improved by adding elitist preserve strategy andremoving duplicates individuals algorithm, and the contact ratio, tooth contact stressand tooth root bending stress were selected as optimal objectives, then optimized themesh parameters of gears.Considering the finite element model was very complicated, Kriging approximatemodel was established on the base of series of finite element analysis for savingcalculation cost, also in order to avoiding calling finite element model to simulate thestress of gears which would lead to mass calculation. The load of gears would changewith the vibration and impaction when during working, so the load dealt with uncertainparameter and described as interval number, and its responsible reliability restrainssolved by the method of non-probability index algorithm. In order to reduce influenceof gears stress variation range which caused by the uncertain parameter, an uncertaintymulti-objective optimization has been run based on interval analysis method, and thencompared with reliability optimization. According to the checking of optimized resultby finite element simulation, that gear stress was lowed and indicated that theoptimized result was effective.