Intelligent Neural Control in Advance Aeropropulsion Engines

Dynacs Engineering Co., Inc. proposes the design and evaluation of an intelligent neural control system in aerospace engines which is capable of withstanding structural failures, component deviations, and unpredictable perturbations. A hybrid connectionist system is advocated to fulfill the critical needs in various operating conditions. The motivation is to pursue a realization of a robust and fault tolerant engine controller with a high degree of autonomy and above acceptable performance. The particular areas that we will address under the proposed efforts will include: identification- dedicated and control-dedicated neural network architectures, real- time learning rules, engine health component assessment, controller association retrieval, and reconfigurable learning control. The Phase 1 research objective is to demonstrate the proof of concept of the proposed neural control approach through numerical simulation of one segment of the overall engine control problem: engine health component assessment for the Pratt & Whitney PW1128 engine model.