基于VMD的注射机熔体温度信号自适应去噪方法研究

Research on Adaptive Denoising Method of Melt Temperature Signal of Injection Molding Machine Based on VMD

为了在含噪声情况下很好的提取注塑过程中注射机监测到的熔体温度信号,提出了一种基于变分模态分解(VMD)的信号自适应去噪方法。在开展去噪之前应当对信号进行预处理,以防止离群点对VMD方法产生影响。之后对预处理过的信号使用VMD方法,按照中心频率的大小来对熔体温度信号进行分解。对信号分解的结果进行含噪情况分析,去除掉含噪较大部分,合并剩余的部分就可以得到去噪后的熔体温度信号。为了验证文中方法的有效性和鲁棒性,首先在模拟熔体温度信号的人造信号中进行实验,结果表明,与其他基于模态分解的方法相比,本实验的方法去噪效果最佳。使用该方法对实际测量的熔体温度信号进行去噪处理,可以很好地去除掉因仪器设备等产生的脉冲噪声,从而得到有效的信息,为后续的建模研究或进行精准控制奠定基础。

In order to extract the melt temperature signal monitored by the injection molding machine in the injection molding process with noise,a signal adaptive denoising method based on variational mode decomposition(VMD)is proposed.This method first preprocesses the signal to prevent outliers from affecting the VMD method.Then use the VMD method on the pre-processed signal to decompose the melt temperature signal according to the center frequency.The results of signal decomposition with noise are analyzed,the larger part of noise is removed,and the denoised melt temperature signal is obtained by combining the remaining parts.In order to verify the effectiveness and robustness of the method in this paper,first,an artificial signal that simulates the melt temperature signal is tested.The results show that the method in this study has the best denoising effect compared with other methods based on modal decomposition.Using this method to de-noise the actual measured melt temperature signal,it can be known that the impulse noise generated by instruments and equipment can be well removed,so as to obtain effective information for subsequent modeling research or lay the foundation for precise control.