Research Activities: 2011
|
A Pilot Study on Diagnostic Sensor Network for Structure Health Monitoring |
|||||
| Status | Complete View Final Report: PDF |
||||
| Sequential Number | R303 | ||||
| Identification Number | 00039451 | ||||
| Matching Research Agency | Missouri University of Science & Technology, Civil Engineering | ||||
| Principal Investigator | Maggie Cheng Associate Professor, Computer Science Missouri University of Science and Technology 325 Computer Science Bldg. Rolla, MO 65401 (573) 341-7501 chengm@mst.edu |
||||
| Student Involvement |
1 Graduate Research Assistant |
||||
Project Objective |
The intent of this study is to provide the multidisciplinary team with a unique opportunity to initiate a new research direction, collect preliminary data, and establish a good track record for collaborative work. Such an effort will allow the team to prepare a competitive proposal for external funds in the following year. The scope of work includes, but are not limited to, (1) Monitor and assess the structural condition of bridges in real-time with sensor networks, and (2) Develop and validate an effective algorithm for the diagnosis of coupled cyber-physical systems in the event of faults using a contour map of structural parameters |
||||
Project Abstract |
This project aims to address the tomography analysis problem in sensor networks. Wireless sensor networks can be placed inside of or on the surface of a structure to estimate the health state of the structure and to detect changes in the structure. For this purpose, the sampled data must be reliable and represent the true state of the structure. In reality, the measured data often come with uncertainty and thus cannot always be interpreted with a theoretical model. When disagreement occurs, the measured data may or may not represent an anomalous state of the structure. Anomaly in data could originate from the source of a structure due to unknown loads and material properties, the error introduced during measurement, and the signal contamination in the process of data communication. One must distinguish the structure anomaly from the sensor network anomaly. For anomaly detection and identification, we must understand the spatial and temporal correlation of the data collected at different sampling points of the structure, and fully understand the network behavior that could introduce anomaly in data. We will provide preliminary results on the asymptotic distribution of test statistics, develop effective means for estimator design and data gathering, and design an efficient algorithm for data processing based on dependence modeling. The ultimate goal is to be able to reconstruct the contour map of some critical parameters of the structure. The algorithm and sensor network will be validated with physical tests of civil engineering structures in laboratory | ||||
Relationship to other Research/Projects |
This is a new line of research for both investigators. Although sensor networks have been used for environment and infrastructure monitoring in the past, but challenging research questions such as data anomaly and network fault have not been addressed adequately. It requires domain knowledge to understand and characterize the spatial and temporal dependence structure of the data in order to detect anomaly and reconstruct the correct contour map of the measurements in strain, cracks, etc. On the other hand, in order to enable a mathematical model for fault diagnosis, the placement of sensors on the infrastructure also needs to be optimized for this purpose. To the best of our knowledge, there is no existing research that has addressed the redundant deployment issue of sensor networks for anomaly detection.
The two investigators have prior working experience in infrastructure health monitoring and sensor networks. However, the proposed research activities are new to both investigators, and can thus create potential opportunities for external funding. |
||||
Transportation-Related Keywords |
Bridge engineering, structural safety, condition assessment, data-driven decision making | ||||
Technology Transfer Activities |
The findings and results will be summarized and presented at the annual UTC conference and other professional conferences as well as the regular CIES meetings. |
||||
Project Deliverables |
The outcomes of this study are expected to publish two or more papers in computer science and civil engineering journals/conferences. These papers will be summarized in a brief final report to fulfill the NUTC requirements at the end of the project. | ||||
Anticipated Benefits |
The main benefits from this study are two folds. First, this study provides an effective way to quantify the uncertainty in structure health monitoring with low-cost and high-accuracy sensor networks. Secondly, this study will result in effective inference algorithms and detectors that can be applied in a wide range of sensor network applications. |
||||
Milestones |
|
||||