| الوصف: |
The applicability of wireless sensor networks (WSNs) has dramatically increased from the era of smart farming and environmental monitoring to the recent commercially successful internet of things (IoT) applications. Simultaneously, diversity in WSN applications has led to the application of specific performance requirements, such as fault tolerance, reliability, robustness and survivability. One important application is structural health monitoring (SHM) in airplanes. Airborne Wireless Sensor Network (AWSN) have received considerable attention in recent times, owing to the many issues that are intrinsic to traditional wire-based airplane monitoring systems, such as complicated cable routing, long wiring, wiring degradation over time, installation overhead, etc. This project examines the SHM of aircraft wing and WSN design (ZigBee), and aspects such as node deployment and power efficient routing, vis-à-vis energy harvesting. Node deployment and power efficient routing protocol are related problems, and so this thesis proposes solutions using optimization techniques for Ant Colony Optimization (ACO), and power transmission profiling using Computer Simulation Technology software (CST). There are three wing models; namely NACA64A410 model, Empty NACA64A410 model for the Wing, and Empty Prismatic model of the wing was specified and simulated in CST software. A simulation was carried out between the frequencies of 100 MHz to 5 GHz, and identified significant variations in the Sij parameter between the frequency range 2.4GHz and 2.5GHz. Critical analysis of the obtained results revealed the presence of a significant impact from wing shape and the wing's inner structure on possible radio wave propagation in the aircraft wing. The different material composition of aircraft wings was also examined to establish the influence of aircraft wing material on radio wave propagation in an aircraft wing. The three materials tested were Perfect Electrical Conductor (PEC), Aluminium, and Carbon Fibre Composites (CFCs). For power transmission profiling (Sij parameter), 130 nodes were deployed in regular and periodic compartments, created by ribs and spars, usually at vantage points and rib openings, so that a direct line of sight could be established. However, four sink nodes were also placed at the wing root, as presented in NC37 and NC38 simulations for aluminium and CFC wing models respectively. The evaluation of signal propagation in aluminium and CFC aircraft wing models revealed CFC wing models allow less transmission than aluminium wing models. A multiple Travelling Salesman (mTSP) problem was formulated and solved, using Ant Colony Optimization in MATLAB to identify optimal topology and optimal routes to support radio propagation in ZigBee networks. Then solving the mTSP problem for different regular deployments of nodes in the wing geometry, it was found that an edgewise communication route was the shortest route for a large number of nodes, wherein 4 fixed sink nodes were placed at the wing root. For a realistic wing model, the different possible configuration of ZigBee units were deduced using rational reasoning, based on results from empty wing models. Besides the determined S-parameter, aircraft wing materials and optimal nodes, the residual energy of each sensor node is also considered an essential criterion to improve the efficiency of ZigBee communications on the aircraft wing. Therefore, a novel hybrid protocol called the Energy-Opportunistic Weighted Minimum Energy (EOWEME) protocol can be formulated and implemented in MATLAB. The comparative results revealed the energy saving of EOWEME protocol is 20% higher compared to the Ad Hoc On-Demand Distance Vector (AODV) routing protocol. However, the need for further energy savings resulted in development of an improved EOWEME protocol when incorporating the clustering concept and the previously determined S-parameter, a number of nodes, and their radiation patterns. Critical evaluation of this improved EOWEME protocol showed a maximum of 10% higher energy savings than the previous EOWEME protocol. To summarize key insights and the results of this thesis, it is apparent that the thesis addresses SHM in aircraft wings, using WSNs from a holistic perspective with the following major contributions, • CST simulations identify power transfer (S-parameter) profiling in various wing models, with no internal structural elements to identify realistic wing with spars, and bars. With an average S-parameter of -107 dB at around 3 m, the communication or transmission range of 1 m was identified to minimize loss of transmitted power. A range less than 1m would cause issues such as interference, reflection etc. • Using a transmission range of 1 m, WSN nodes were assessed for shortest route commensurate with energy efficient packet transmission to sink node from the farthest node; i.e. near the wing tip. The shortest routes converged to travel along the length of the wing in the case of an empty wing model, however it was also observed in a realistic wing model, where internal structural elements constrained node deployment. An average distance of nearly 13 m required data transmitted from the farthest nodes to reach the sink nodes. Increasing the nodes however increased the distance required to up to 20 m in the case of 240 nodes. • A new routing protocol, EOWEME was formulated, showing 20% greater energy savings than AODV in the realistic wing model. |
