Optimized synthesis of cost‐effective, controllable oil system architectures for turbofan engines

Abstract

Turbofan oil systems are used to provide lubrication and cooling in the engine . There is an increasing interest in oil system architectures which utilize electric pumps and/or valves to give optimized control of flows to individual oil chambers, leading to improved thermal management of oil and lubrication efficiency. The challenges here lie in the trade‐off between increasing controllability and minimizing the addition of new components, which adds unwanted production and maintenance costs. This paper formulates the oil system architecture design as a constrained, multiobjective optimization problem. An architecture is described using a graph with nodes representing components and edges representing interconnections between components. A fixed set of nodes called the architecture template is provided as an input and the edges are optimized for a multicriteria objective function. A heuristic method for determining similarities between the different oil chamber flow requirements is presented. This is used in the optimization to evaluate the controllability objective based on the structure of the valve architecture. The methodology provides benefits to system designers by selecting cheaper architectures with fewer valves when the need to control oil chambers separately is small. The effect of manipulating the cost/controllability criteria weightings is investigated to show the impact on the resulting architecture.

Utilizing a maximin optimization approach to maximize system resiliency

“…8 This paper describes how to apply a maximin optimization approach to the design of a system for a wide range of operating environments and demonstrates the technique using an air defense system of systems. While mathematical programing is frequently used in system design, [9][10][11] it is typically used to optimize the system against a single or small number of operating environments. 1,12 The technique described here evolves the operating environment as the system design changes, to prevent the system from being designed only for a relatively small percentage of the systems desired operating space.…”

“…There has been a significant amount of work focused on applying mathematical programing techniques to system design. [9][10][11] A very common approach is in increasing the reliability of a system by designing redundant components and using Mixed Integer Non-Linear Programming (MINLP) to determine the reliability and number required of each component. 25 While MINLP has many advantages, including speed and an optimal solution for linear problems, it becomes restrictive in all but the simplest system design problems.…”