الفهرس 
IoT Elements
Smart Utilities Traffic Scheduling Algorithm (SUTSA)Only 14 pages are availabe for public view
 |  | from | 150 |  |  |
| | | from | 150 | | |
Abstract
The Internet of Things (IoT) is an ecosystem that connects billions of smart devices, meters, and sensors. These devices and sensors collect and share data for use and evaluation by organizations in different industry sectors. Humans may use the IoT to live and work more intelligently and gain total control over their lives. Consequently, IoT can be used to connect devices and integrate them with new digital technologies for customers. On the other hand, smart utility companies in electric, gas, and water sectors need to use upgraded power system components to replace the old electrical infrastructure with the smart grid.
In addition, load forecasting is one of the main concerns for power utility companies. It plays a significant role in planning decisions, scheduling, operations, pricing, customer satisfaction, and system security. This helps smart utility companies deliver services more efficiently and analyze their operations in a way that can help optimize performance, detect growing problems in real-time, and initiate fixes to avoid unplanned service interruptions. The electricity load can be predicted using several machine-learning methods.
This thesis aims at using the power line network as an infrastructure for the smart grid and using different machine learning algorithms to analyze the data collected from the smart meter. This was achieved by means of the following steps:
First, a new Smart Utilities Traffic Scheduling Algorithm (SUTSA) for smart meter data collection, including readings and alarms, is proposed. This algorithm utilizes the current narrowband PL network, which alleviates the need to develop a new infrastructure dedicated to smart meter data collection. It has three different models for collecting data that are suitable for different practical cases. Second, the collected meter readings are analyzed using Multiple Linear Regression (MLR), Random Forests (RF), Artificial Neural Networks (ANN), the Chi-square Automatic Interaction Detector (CHAID), Xtreme Gradient Boosting (XGBoost), and an Automatic Regression Integrated Moving Average (ARIMA) to provide valuable insight into consumers’ actions and desires. It also helps smart utility companies with load forecasting, upgrade planning decisions, scheduling, operations, pricing, customer satisfaction, and system security.
The thesis also focuses on enhancing electricity consumption predictions in Middle Eastern countries. The lack of research that focuses on the consumption behavior of this region results in poor electricity load forecasting.
To validate the accuracy of the proposed models, a simulation was performed using OPNET Modeler 14.5. The results proved that the proposed model achieved full network bandwidth utilization in different situations based on application requirements. Also, the machine learning algorithms are implemented and evaluated using the IBM SPSS Modeler, IBM SPSS statistics, and Python code.
In conclusion, the contributions proposed in this thesis will help to improve the use of PLC as an infrastructure network for smart meters. In addition, the research results can assist electricity companies with load forecasting, malfunction detection, and electricity theft detection, noticing the shortage or excess in electricity production.