基于熵值法径向基神经网络的清扫车吸尘口垃圾颗粒驻留时间预测

针对清扫车吸尘口内垃圾颗粒驻留时间受到结构类、固相类及气相类因素多类型、非线性作用预测难的问题，提出了基于熵值法的径向基神经网络(RBFNN)的垃圾颗粒驻留时间预测方法,该方法考虑了吸尘负压、滚刷转速、颗粒质量、颗粒密度、颗粒流量和吸尘管直径6种因素，将采用熵值法求解的因素权重作为输入扰动变量，建立了垃圾颗粒驻留时间的预测模型。结果表明：与采用传统RBFNN的预测方法相比，所提方法具有预测精度高的优点，可较好地解决清扫车吸尘口垃圾颗粒驻留时间预测难的问题，有助于提升清扫车吸尘系统设计水平。

Residence Time Prediction of Waste Particles in Dust Suction Ports for Sweepers in RBFNN Based on Entropy Method

For the problems that the residence time of waste particles in the dust suction ports for the sweepers was difficult to predict due to multi-type and non-linear interaction among structural, solid phase and gas phase factors, the residence time prediction method of waste particles was proposed in RBFNN based on entropy method, which considered six factors including suction vacuum negative pressure, roller rotation speed， particle mass, particle density, particle flow rate, and diameter of the suction tubes. The factor weights solved by the entropy method were used as the input disturbance variables, and the prediction model of the residence time of waste particles was established. The results show that compared with the traditional RBFNN prediction method, the proposed method has the advantage of high prediction accuracy, and may solve the problem of difficult prediction of residence time of waste particles in the dust suction ports for the sweepers, which is helpful to improve the design level of the dust suction systems for the sweepers.