改进的模糊化神经网络的土壤振动掘削阻力软测量模型

    Soft-sensing model on vibration cutting resistance from rock and soil based on improved fuzzy neural networks theory

    • 摘要: 为实现土壤振动掘削过程节能减阻,分析了振动频率、振幅和插入速度等对土壤振动掘削阻力的影响,并以振动频率、振幅和插入速度等特征参数作为二次变量,采用清晰集构造模糊集方法对土壤振动掘削阻力软测量模型特征参数进行模糊化,利用改进的模糊化神经网络建立了土壤振动掘削阻力软测量模型。土壤振动切削力软测量实际应用结果表明,土壤振动掘削阻力软测量值的建模精度和泛化能力是很高的,所得的最大训练相对误差约为0.67%,最小测试相对误差约为-0.4%,有利于土壤振动掘削阻力的快速精确测量。

       

      Abstract: In order to save energy and reduce cutting resistance in the process of vibratory excavating for rock and soil, the impacts of vibration frequency, amplitude and inserting velocity on vibratory excavating resistance from rock and soil were analyzed. Based on the method for constructing a fuzzy set by using clear sets, a soft-sensing model on vibration cutting resistance from rock and soil was brought up by using of the vibration frequency, amplitude and inserting velocity as accessorial parameters. The application results revealed that the soft sensing model was of much high modeling precision and generalization capability and its most train relative error was about 0.67%, its least test relative error was about -0.4%. And the soft-sensing model is very useful for measurement of vibration cutting resistance from rock and soil quickly and precisely.

       

    /

    返回文章
    返回