Research on terminal traffic prediction and dynamic security assessment method for smart power plants based on the ARIMA model

This study utilizes the ARIMA (AutoRegressive Integrated Moving Average) model to model and analyze terminal data traffic in smart plant networks, proposing a traffic prediction and dynamic threshold determination method based on the ARIMA model. The method employs differencing to transform non-stationary data into stationary sequences and combines the ARMA (AutoRegressive Moving Average) model for traffic trend prediction. The root mean square error (RMSE) is used to evaluate the model's fitting performance, ensuring the accuracy of the prediction results. By adopting a dynamic threshold determination strategy based on probabilistic upper bounds, a real-time trust evaluation and anomaly detection mechanism for zero-trust environments is established. When traffic exceeds the threshold range, the system can promptly issue alerts and adjust access permissions. Compared to traditional traffic prediction methods, which often lack robustness in handling the high volatility and non-stationarity of smart plant network traffic, this study introduces a method combining ARIMA modeling with probabilistic upper bound thresholds, offering greater adaptability and reliability for real-time applications. Experimental results demonstrate that this method performs with high reliability in capturing the fluctuating characteristics of network traffic and improving the accuracy of anomaly detection, providing effective technical support for the security of smart plant networks.