Abstract It is important to concentrate on monitoring and control of the pipeline transportation system before the failure resulting in fatal accidents. To enhance the supervision performances, the SCADA (Supervisory Control and Data Acquisition) platform is incorporated with IoT by utilizing the NB-IOT module holding a high-level engineering interface. In the proposed methodology, SCADA with the LQR-PID controller serves as Local Intelligence. When the local intelligence fails to react proactively during risk occurrences, immediately its performance is deactivated by the webserver through the NB (Narrow Band)-IoT module. For experimental real-time validation of the proposed work, a lab-scale DCS (Distributed Control System) based fluid transportation system is undertaken where flow and pressure prevail to be the most influencing parameters during risk occurrences in the pipelines. Also, the performance analyses are validated experimentally using unsupervised K-means clustering to identify abnormality caused by blockage and crack in the pipeline on the cloud-stored data.

中文翻译：

---

使用机器学习进行基于物联网的流体运输监测和控制

摘要 在管道运输系统发生故障导致致命事故之前，对管道运输系统进行监测和控制是非常重要的。为了提高监管性能，SCADA（监管控制和数据采集）平台通过利用具有高级工程接口的 NB-IOT 模块与物联网相结合。在建议的方法中，带有 LQR-PID 控制器的 SCADA 用作本地智能。当本地智能在风险发生期间未能主动做出反应时，网络服务器会立即通过 NB（窄带）-IoT 模块停用其性能。对于拟议工作的实验实时验证，一个实验室规模的基于 DCS（分布式控制系统）的流体输送系统被采用，其中流量和压力在管道中发生风险时是影响最大的参数。此外，性能分析使用无监督 K 均值聚类进行实验验证，以识别由云存储数据上的管道堵塞和裂缝引起的异常。