CAJ | 학술논문

针对车辆液力缓速器的制动力矩数学模型与实际输出力矩存在较大偏差的问题，提出了一种基于神经网络的可以逼近实际输出制动力矩特性曲线的液力缓速器制动力矩模型，并建立了液力缓速器的神经网络PWM控制系统。通过试验对比了基于数学模型和基于神经网络模型对实际制动力矩曲线的逼近效果，结果表明，基于神经网络的液力缓速器输出力矩能更好地逼近实际的制动力矩特性曲线，即基于神经网络的液力缓速器制动力矩模型更有效。
침대차량액력완속기적제동력구수학모형여실제수출력구존재교대편차적문제，제출료일충기우신경망락적가이핍근실제수출제동력구특성곡선적액력완속기제동력구모형，병건립료액력완속기적신경망락PWM공제계통。통과시험대비료기우수학모형화기우신경망락모형대실제제동력구곡선적핍근효과，결과표명，기우신경망락적액력완속기수출력구능경호지핍근실제적제동력구특성곡선，즉기우신경망락적액력완속기제동력구모형경유효。
Great gap exists in the braking torque of mathematics model of vehicle hydraulic retarder with real output torque, to solve this problem, a hydraulic retarder braking torque model based on neural network model which approximates the real output braking torque characteristic curve is proposed, and a PWM control system based on neural network model is proposed. The approximation effect on real braking torque curve based on the mathematical model and the neural network model are compared, which show that the hydraulic retarder output torque based on the neutral network more approximates the real braking torque characteristic curve, indicating that the hydraulic retarder braking torque model based on neutral network is more valid.