A FORTRAN, MS-DOS operating system algorithm was constructed and implemented to predict ester-based lubricant behavior in a turbine engine. Input data consists of selected lubricant flow rates, bulk lubricant and bearing temperatures, and laboratory generated data. Execution of the program uses lubricant property data as a function of time at several temperatures to produce mathematical functions describing total acid number, viscosity change and evaporation as a function of temperature at selected times. Static coker data are employed to estimate coking at the hot end bearing cover. The program output data are predicted values of critical lubricant properties as a function of engine hours, reflecting lubricant additions to the system because of evaporative and seal losses and also degradation predicted to occur during cool-down cycles. Computer-generated data are compared with actual engine data; the benefits and limitations of the program are identified.Presented at the 43rd Annual Meeting in Cleveland, Ohio May 9–12, 1988