Accuracy metrics have been widely used for the evaluation of predictions in machine learning. However, the selection of an appropriate accuracy metric for the evaluation of a specific prediction has not yet been specified. In this study, seven of the most used accuracy metrics in machine learning were summarized, and both their advantages and disadvantages were studied. To achieve this, the acoustic emission data of damage locations were collected from a pile hit test. A backpropagation artificial neural network prediction model for damage locations was trained with acoustic emission data using six different training algorithms, and the prediction accuracies of six algorithms were evaluated using seven different accuracy metrics. Test results showed that the training algorithm of “TRAINGLM” exhibited the best performance for predicting damage locations in deep piles. Subsequently, the artificial neural networks were trained using three different datasets collected from three acoustic emission sensor groups, and the prediction accuracies of three models were evaluated with the seven different accuracy metrics. The test results showed that the dataset collected from the pile body-installed sensors group exhibited the highest accuracy for predicting damage locations in deep piles. Subsequently, the correlations between the seven accuracy metrics and the sensitivity of each accuracy metrics were discussed based on the analysis results. Eventually, a novel selection method for an appropriate accuracy metric to evaluate the accuracy of specific predictions was proposed. This novel method is useful to select an appropriate accuracy metric for wide predictions, especially in the engineering field.

The acoustic emission (AE) technique has been widely used to monitor damage for concrete structures in underground space. To analyze the damage status successfully, it is important to understand the characteristics of AE parameters during failure for concrete-based-underground structures. The AE parameters are related to the target material characteristics (i.e., sand content ratio) and failure steps (i.e., stress levels) of concrete specimens. In this study, the concrete specimens were prepared with different sand content ratio. Uniaxial compressive tests were performed for the specimens with AE monitoring to investigate the characteristics of AE parameters in the time domain (i.e., count, energy, and amplitude) and frequency domain (i.e., initial frequency and peak frequency). The experimental results were analyzed with the sand content ratio and failure step (1–4) in order to determine the suitable AE parameter as a monitoring index. The most suitable parameter was shown as the cumulative energy parameter of AE with high sensitivity.

As one of the non-destructive testing (NDT) methods, acoustic emission (AE) can be widely applied to the field of engineering and applied science owing to its advantageous characteristics. In particular, the AE method is effectively applied to monitor concrete structures in civil engineering. For this technology to be employed in a monitoring system, it is necessary to investigate the propagation characteristics of the AE in structures. Hence, this study investigates the characteristics of AE in concrete structures to evaluate the field applicability of AE monitoring systems. To achieve this goal, experiments employing an AE system are conducted for concrete structures 20 × 0.2 × 1.2 m in length, width, and height, respectively, to explore the AE parameters according to the impact energy. Among all AE parameters, absolute energy is determined to be most sensitive factor with respect to the impact energy. In addition, the attenuation effect of the AE wave is quantitatively evaluated according to the wave propagation distance. Moreover, the concept of effective distance is newly suggested based on the experimental results. The effective distance is shown to increase as the impact energy increases, although the increased effective distance is limited because the damaged AE signal is of high frequency. This study helps improve the field applicability of AE monitoring systems by suggesting suitable AE sensor spacing, which contributes to promote the practice of technology.

This study presents a cost-based optimization model for the design of isolated foundations in cohesive soils. The optimization algorithm not only incorporates safety requirements in the form of ultimate limit state (ULS) and serviceability limit state (SLS) criteria but also deals with the economics simultaneously. In that regard, the generalized reduced gradient (GRG) method is used for the optimization purpose to achieve the least construction cost of an isolated foundation along with the integration of design parameters as optimization variables. The optimization technique is elaborated using a design example in silty clayey soil and the results of the optimized design are compared with those of the conventional design. The optimization model shows that the optimized design can reduce the construction cost by up to 44% as compared to the conventional design cost for the particular example. Moreover, a sensitivity analysis is also performed to evaluate the quantitative impact of cohesive soil properties, design load, and groundwater table on the construction cost. The results indicate that the construction cost majorly depends on the combined effect of four key parameters: Young’s modulus, recompression index, design load, and groundwater table.