摘要
介绍了某研究项目中所使用的液力偶合器数学建模新方法———人工神经网络方法.从舰船动力装置动态建模与仿真的实际需求,分析了液力偶合器数学建模的困难之处:寻找一种通用的数学形式,以实现:①额定工况滑差及其附近实验数据的内插与外延;②过渡工况(如液力偶合器结合时其滑差从100%变到额定工况滑差附近的过程中)液力偶合器数学模型的设计.叙述了液力偶合器的人工神经网络数学建模的要点,分析了液力偶合器人工神经网络数学模型在舰船动力装置计算机仿真中的结果.最后,总结了液力偶合器人工神经网络数学模型今后需进一步完善的若干方面.
A new approach of building a mathematical model of ship's fluid coupling, i.e., the Artificial Neural Network approach, is introduced in the paper. First, from the requirement of dynamic modeling and computer simulation of ship power plant, a tough task which should be tackled in building a mathematical model of fluid coupling is analyzed, i.e., this model should have universal form and be used for ① performing interpolation and extrapolation of test data around the area of the rated working condition; ② calculating the transient process, for example, the engagement/disengagement of fluid coupling. Then, the key points needed in application of the Artificial Neural Network (ANN) to building a mathematical model of fluid coupling are stated, some results of computer simulation of fluid coupling's ANN model in ship power plant are analyzed. Last, some aspects, which should be done for further improving the mathematical model, are predicted.
出处
《海军工程大学学报》
CAS
2002年第6期40-45,共6页
Journal of Naval University of Engineering
关键词
数学模型
神经网络
仿真
液力偶合器
动力装置
舰船
mathematical model
neural network
computer simulation
fluid coupling
power plant
ship