The time-delay in a networked control system is difficult to predict accurately. In order to improve prediction performance, a time-delay prediction method is proposed. Firstly, the approximation components with low-frequency and approximation components with a high frequency of time-delay sequence are obtained through the decomposition and reconstruction of Db3 (Daubechies 3) wavelet transform. Secondly, the characteristics of the approximate component and detail components are analyzed. Fractal autoregressive integrated moving average (FARIMA) is chosen as the prediction model for an approximate component. Autoregressive integrated moving average (ARIMA) is chosen as the prediction model for detail components. Finally, the final prediction value can be obtained by the prediction value of FARIMA adding the prediction value of ARIMA. The proposed prediction method combines the advantages of FARIMA and ARIMA. The case study results show that the proposed prediction method has better prediction accuracy.

网络化控制系统中的时延很难准确预测。为了提高预测性能，提出了一种时延预测方法。首先，通过对Db3(Daubechies 3)小波变换进行分解重构，得到时延序列的低频近似分量和高频近似分量。其次，分析了近似分量和细部分量的特点。选择分形自回归积分移动平均 (FARIMA) 作为近似分量的预测模型。选择自回归综合移动平均（ARIMA）作为细节分量的预测模型。最后，通过 FARIMA 的预测值加上 ARIMA 的预测值可以得到最终的预测值。所提出的预测方法结合了 FARIMA 和 ARIMA 的优点。案例研究结果表明，所提出的预测方法具有较好的预测精度。