Crack monitoring effectiveness of CPGs and EFSs depends on the direct alignment of the sensor with the direction of crack propagation. Additionally, proper bonding of the CPG is crucial to avoid the situation where the crack in the substrate propagates without breaking the resistance strands or the resistance strands break before the crack passes through it, resulting in measurement errors [14]. Flexibility and durability are concerns for the EFS, which uses an electrolyte that eventually dries out. Furthermore, the EFS uses a purely empirical algorithm to detect fatigue crack initiation [27,28], which may only apply for the specific situations in which it has been evaluated. Finally, mounted at an existing crack tip, the electrochemical fatigue crack sensor is small and designed to detect crack initiation; it is not suitable for monitoring crack propagation. Eddy current-based sensors can detect microcracks in metallic structures using proximity sensing [29]. MWMs (using eddy currents) can reliably detect buried and surface cracks. An MWM-array sensor is most effective when the crack is underneath the sensor [30]. The MWM-array sensor system uses a measurement technique that may require expensive hardware and specialized software to acquire and interpret the raw data for crack monitoring.
Fractographic approach to validate calculated crack length using BFS data. The figure shows (a) a plot of crack length vs. number of load cycles with (b) an optical image showing the corresponding beach marks directly observed on the fracture surface. The high-magnification SEM micrographs (insets 1 and 2) show locations 1 and 2 of fatigue striations with different spacings (red arrows in each inset point to an individual striation).
Mark A. Weiss, Data Structures and Algorithm Analysis in C 4th crack
RvNN can achieve predictions in a hierarchical structure also classify the outputs utilizing compositional vectors [57]. Recursive auto-associative memory (RAAM) [58] is the primary inspiration for the RvNN development. The RvNN architecture is generated for processing objects, which have randomly shaped structures like graphs or trees. This approach generates a fixed-width distributed representation from a variable-size recursive-data structure. The network is trained using an introduced back-propagation through structure (BTS) learning system [58]. The BTS system tracks the same technique as the general-back propagation algorithm and has the ability to support a treelike structure. Auto-association trains the network to regenerate the input-layer pattern at the output layer. RvNN is highly effective in the NLP context. Socher et al. [59] introduced RvNN architecture designed to process inputs from a variety of modalities. These authors demonstrate two applications for classifying natural language sentences: cases where each sentence is split into words and nature images, and cases where each image is separated into various segments of interest. RvNN computes a likely pair of scores for merging and constructs a syntactic tree. Furthermore, RvNN calculates a score related to the merge plausibility for every pair of units. Next, the pair with the largest score is merged within a composition vector. Following every merge, RvNN generates (a) a larger area of numerous units, (b) a compositional vector of the area, and (c) a label for the class (for instance, a noun phrase will become the class label for the new area if two units are noun words). The compositional vector for the entire area is the root of the RvNN tree structure. An example RvNN tree is shown in Fig. 5. RvNN has been employed in several applications [60,61,62].
The resulting data may either be fit as is, accounting for the observed smearing, or desmeared before fitting using one of several available algorithms (e.g. Kline 2006; Ilavsky and Jemian 2009). The latter makes association with SANS results at higher Q, and results obtained from image analysis at lower Q easier, but may introduce additional noise and uncertainties, especially if the scattering pattern is not circular.
It is far beyond the scope of this review to even begin an analysis of the applications of image analysis to geological samples. However, in the context of analyzing and quantifying pore structures some discussion is appropriate, because analysis of low-magnification SEM/BSE or X-ray computed tomographic images can be used to extend the scale range analyzed by SAS experiments, and thus imagery can be used for pore characterization beyond that provided by point counting. In addition, in the process of obtaining and processing these data one generates binary images of the pore structure of the rock, typically at scales greater than approximately 1 mm that can then be used to provide further quantification of the pores structure at these scales using other statistical techniques that require the two- or three-dimensional data available in the images themselves.
As has been noted above, SAS data suggest that pore structures in rocks exhibit both surface and mass fractal behavior. While the scattering data do not directly show what those structures look like, as noted above structure and form factor models such as those suggested by Beaucage (1995, 1996) and Beaucage et al. (1995, 2004) are based on models of this structure. Imaging data provides the opportunity to extend this analysis to a consideration of direct box-counting fractal (Block et al. 1990) and multifractal behavior based on actual observations.
2ff7e9595c
Comments