On Self-organising Diagnostics in Impact Sensing Networks

Prokopenko, Mikhail; Wang, Peter; Scott, Andrew; Gerasimov, Vadim; Hoschke, Nigel; Price, D. C.

doi:10.1007/11554028_25

Cited by 4 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For this reason, the study conducted by Omenzetter formulated a vector seasonal autoregressive integrated moving average (ARIMA) model for the recorded strain data through which unusual events as well as structural change or damage sustained by the structure could be revealed [11]. In addition to these methods, some artificial intelligence methods such as neural networks, fuzzy processing, genetic algorithms and combinations of these methods have been proposed to handle such problems [12][13][14][15]. However, real-time assessment and low data storage requirements are still expected for any practical solution.…”

Section: Introductionmentioning

confidence: 99%

A bio-inspired memory model for structural health monitoring

Zheng¹,

Zhu²

2009

Meas. Sci. Technol.

View full text Add to dashboard Cite

Long-term structural health monitoring (SHM) systems need intelligent management of the monitoring data. By analogy with the way the human brain processes memories, we present a bio-inspired memory model (BIMM) that does not require prior knowledge of the structure parameters. The model contains three time-domain areas: a sensory memory area, a short-term memory area and a long-term memory area. First, the initial parameters of the structural state are specified to establish safety criteria. Then the large amount of monitoring data that falls within the safety limits is filtered while the data outside the safety limits are captured instantly in the sensory memory area. Second, disturbance signals are distinguished from danger signals in the short-term memory area. Finally, the stable data of the structural balance state are preserved in the long-term memory area. A strategy for priority scheduling via fuzzy c-means for the proposed model is then introduced. An experiment on bridge tower deformation demonstrates that the proposed model can be applied for real-time acquisition, limited-space storage and intelligent mining of the monitoring data in a long-term SHM system.

show abstract