Andrew T. Zimmerman

A Parallel Simulated Annealing Architecture for Model Updating in Wireless Sensor Networks

In recent years, wireless sensing technologies have provided a much sought-after alternative to expensive cabled monitoring systems. Wireless sensing networks forego the high data transfer rates associated with cabled sensors in exchange for low-cost and low-power communication between a large number of sensing devices, each of which features embedded data processing capabilities. As such, a new paradigm in large-scale data processing has emerged; one where communication bandwidth is somewhat limited but distributed data processing centers are abundant. By taking advantage of this grid of computational resources, data processing tasks once performed independently by a central processing unit can now be parallelized, automated, and carried out within a wireless sensor network. By utilizing the intelligent organization and self-healing properties of many wireless networks, an extremely scalable multiprocessor computational framework can be developed to perform advanced engineering analyses. In this study, a novel parallelization of the simulated annealing stochastic search algorithm is presented and used to update structural models by comparing model predictions to experimental results. The resulting distributed model updating algorithm is validated within a network of wireless sensors by identifying the mass, stiffness, and damping properties of a three-story steel structure subjected to seismic base motion.

Hybrid wireless hull monitoring system for naval combat vessels

There is increasing interest by the naval engineering community in permanent monitoring systems that can monitor the structural behaviour of ships during their operation at sea. This study seeks to reduce the cost and installation complexity of hull monitoring systems by introducing wireless sensors into their architectural designs. Wireless sensor networks also provide other advantages over their cable-based counterparts such as adaptability, redundancy, and weight savings. While wireless sensors can enhance functionality and reduce cost, the compartmentalised layout of most ships requires some wired networking to communicate data globally throughout the ship. In this study, 20 wireless sensing nodes are connected to a ship-wide fibre-optic data network to serve as a hybrid wireless hull monitoring system on a high-speed littoral combat vessel (FSF-1 Sea Fighter). The wireless hull monitoring system is used to collect acceleration and strain data during unattended operation during a one-month period at sea. The key findings of this study include that wireless sensors can be effectively used for reliable and accurate hull monitoring. Furthermore, the fact that they are low-cost can lead to higher sensor densities in a hull monitoring system thereby allowing properties, such as hull mode shapes, to be accurately calculated.

Market-based resource allocation for distributed data processing in wireless sensor networks

In recent years, improved wireless technologies have enabled the low-cost deployment of large numbers of sensors for a wide range of monitoring applications. Because of the computational resources (processing capability, storage capacity, etc.) collocated with each sensor in a wireless network, it is often possible to perform advanced data analysis tasks autonomously and in-network, eliminating the need for the post-processing of sensor data. With new parallel algorithms being developed for in-network computation, it has become necessary to create a framework in which all of a wireless networks scarce resources (CPU time, wireless bandwidth, storage capacity, battery power, etc.) can be best utilized in the midst of competing computational requirements. In this study, a market-based method is developed to autonomously distribute these scarce network resources across various computational tasks with competing objectives and/or resource demands. This method is experimentally validated on a network of wireless sensing prototypes, where it is shown to be capable of Pareto-optimally allocating scarce network resources. Then, it is applied to the real-world problem of rupture detection in shipboard chilled water systems.

Market-based computational task assignment within autonomous wireless sensor networks

In recent years, improved wireless technologies have enabled the low-cost deployment of large numbers of sensors for a variety of applications across different engineering disciplines. Because of the computational resources (processing capability, storage capacity, etc.) distributed throughout these sensing networks, it is often possible to perform advanced data analysis tasks autonomously and in-network, eliminating the need for the post-processing of sensor data. With new parallel algorithms being developed for in-network computation, it has become necessary to create a framework in which the computational resources available throughout a wireless sensing network can be best utilized in the midst of competing computational requirements. In this study, a Pareto-optimal market-based method is developed in order to autonomously distribute various computational tasks with competing objectives and/or resource demands across available network resources. This method is experimentally validated on a network of wireless sensing prototypes.

Decentralized agent-based control of chilled water plants using wireless sensor and actuator networks

The resiliency of a ship is dependent upon the resiliency of the various engineering plants that operate the ship. Especially for combatant ships, engineering plants must be reconfigurable when damage occurs to ensure the ship has fight-through capabilities. Furthermore, reduced manning on ships necessitates the automated operation of engineering plants, especially their reconfiguration during times of battle damage. Wireless telemetry has been proposed in lieu of traditional tethered architectures for the monitoring and control of shipboard engineering plants. In this study, wireless nodes capable of sensing and actuation are explored for the automated control of a chilled water plant. A utility oriented agent-based control network is proposed as a scalable and robust approach to the automated configuration of a chilled water plant. To illustrate the performance of the proposed control and reconfiguration architecture, a small-scale chilled water demonstrator is utilized. A network of wireless sensing and actuation nodes are shown to be highly effective in monitoring and reconfiguring the chilled water plant under varying operational conditions to achieve its operational objectives.

A Framework for Embedded Load Estimation from Structural Response of Wind Turbines

The international push in the development of energy that is sustainable in the long term is driving technological improvements in the area of wind-generated energy. Pushing the limits of current knowledge, turbines now feature increasingly slender towers, larger gear boxes, and significantly longer blades in search of greater capacities and improved efficiency. In addition, siting concerns are leading planners to build these structures in increasingly challenging environments where they are subject to harsh and poorly characterized loadings (particularly in off-shore applications where wind and wave interactions are poorly understood). Future safe and economical designs require accurate characterization of design loads, however direct measurement of wind loads on turbines can be problematic due to the disturbance caused by the wind’s interaction with the turbine blades. This paper presents a novel means of estimating wind loading from the dynamic response of the turbine tower to these loads. A model of the structure is derived using the assumed modes method and then updated using dynamically collected acceleration data a there the input-output relationships are established and input loading spectra estimated. The method relies on reduced-order modal space models making it suitable for real-time operation or embedment in a low-cost autonomous (perhaps wireless) monitoring system. Results derived for a full scale structure under lateral seismic loading are presented.

Automated wind load characterization of wind turbine structures by embedded model updating

The continued development of renewable energy resources is for the nation to limit its carbon footprint and to enjoy independence in energy production. Key to that effort are reliable generators of renewable energy sources that are economically competitive with legacy sources. In the area of wind energy, a major contributor to the cost of implementation is large uncertainty regarding the condition of wind turbines in the field due to lack of information about loading, dynamic response, and fatigue life of the structure expended. Under favorable circumstances, this uncertainty leads to overly conservative designs and maintenance schedules. Under unfavorable circumstances, it leads to inadequate maintenance schedules, damage to electrical systems, or even structural failure. Low-cost wireless sensors can provide more certainty for stakeholders by measuring the dynamic response of the structure to loading, estimating the fatigue state of the structure, and extracting loading information from the structural response without the need of an upwind instrumentation tower. This study presents a method for using wireless sensor networks to estimate the spectral properties of a wind turbine tower loading based on its measured response and some rudimentary knowledge of its structure. Structural parameters are estimated via model-updating in the frequency domain to produce an identification of the system. The updated structural model and the measured output spectra are then used to estimate the input spectra. Laboratory results are presented indicating accurate load characterization.

Distributed data processing within dense networks of wireless sensors using parallelized model updating techniques

As costs associated with wireless sensing technologies continue to decline, it has become feasible to deploy dense networks of tens, if not hundreds of wireless sensors within a single structural system. Additionally, many state-of-the-art wireless sensing platforms now integrate low-power microprocessors and high-precision analog-to-digital converters in their designs. As a result, data processing tasks can be efficiently distributed across large networks of wireless sensors. In this study, a parallelized model updating algorithm is designed for implementation within a network of wireless sensing prototypes. Using a novel parallel simulated annealing search method optimized for in-network execution, this algorithm efficiently assigns model parameters so as to minimize differences between an analytical model of the structure and wirelessly collected sensor data. Validation of this approach is provided by updating a lumped-mass shear structure model of a six-story steel building exposed to seismic base motion.

Parallelized Simulated Annealing for Model Updating in Ad-Hoc Wireless Sensing Networks

The engineering community has recently begun to adopt wireless sensing technologies for use in many sensing applications. These low-cost sensors provide an optimal setting for dense sensing networks, and can make large amounts of sensor data available. Also, the computational power embedded within each sensing node allows a wireless network to interrogate data within the network and in real-time. In a monitoring situation, these capabilities can be leveraged to detect and locate changes in system properties that could be an early indication of malfunction within a complex physical system. In this paper, a distributed model updating technique is embedded within the computational core of a wireless sensing network. Using a novel distributed implementation of the simulated annealing method, an ad-hoc network of wireless sensing units can determine updated system properties by iteratively matching data derived from an analytical model of the system with collected sensor data. By comparing updated system properties with those obtained in a baseline state, a wireless sensing network can detect and locate changes in the system. Experimental verification of this technique is provided on a network of wireless sensor prototypes using simulated results from a Navy ship-board chilled water system monitored with a wireless sensor network.

Automated mode shape estimation in agent-based wireless sensor networks

Recent advances in wireless sensing technology have made it possible to deploy dense networks of sensing transducers within large structural systems. Because these networks leverage the embedded computing power and agent-based abilities integral to many wireless sensing devices, it is possible to analyze sensor data autonomously and in-network. In this study, market-based techniques are used to autonomously estimate mode shapes within a network of agent-based wireless sensors. Specifically, recent work in both decentralized Frequency Domain Decomposition and market-based resource allocation is leveraged to create a mode shape estimation algorithm derived from free-market principles. This algorithm allows an agent-based wireless sensor network to autonomously shift emphasis between improving mode shape accuracy and limiting the consumption of certain scarce network resources: processing time, storage capacity, and power consumption. The developed algorithm is validated by successfully estimating mode shapes using a network of wireless sensor prototypes deployed on the mezzanine balcony of Hill Auditorium, located on the University of Michigan campus.