The knowledge on strength properties of porous metals in compression is essential in tailored application design, as well as in elaboration of general material models. In this article, the authors propose specification details of the ANN architecture for adequate modelling of the phenomenon of compressive behaviour of open-cell aluminum. In the presented research, an algorithm was used to build different structures of artificial neural networks (ANNs), which approximated stress-strain relations of an aluminum sponge subjected to compression. Next, the quality of the built approximations was appraised. The mean absolute relative error (MARE), coefficient of determination between outputs and targets R2, root mean square error (RMSE), and mean square error (MSE) were assumed as criterial measures for the assessment of the fitting quality. The studied neural networks (NNs) were two-layer feedforward networks with different numbers of neurons in the hidden layer. A set of experimental stress-strain data from quasistatic uniaxial compression tests of open-cell aluminum of various apparent densities was used as data for training of neural networks. Analysis was performed in two modes: in the first one, all samples were taken for training, and in the second case, one sample was left out during training in order to play the role of external data for testing the trained network later. The taken out samples were maximum and minimum density samples (for extrapolation) and one random from within the density interval. The results showed that good approximation on the engineering level MARE<5% was reached for teaching networks with ≥7 neurons in the hidden layer for the first studied case and with ≥8 neurons for the second. Calculations on external data proved that 8 neurons are enough to actually obtain MARE<10%. Moreover, it was shown that the quality of approximation can be significantly improved to MARE≈7% (tested on external data) if the initial region of the stress-strain relation is modelled by an additional network.

It is a common situation that seismic excitations may lead to collisions between adjacent civil engineering structures. This phenomenon, called earthquake-induced structural pounding, may result in serious damage or even the total collapse of the colliding structures. Filling the gap between two buildings erected close to one another by using visco-elastic materials can be considered to be one of the most effective methods to avoid seismic pounding. In this paper, a new polymer–metal composite material made of polyurethane and closed-cell aluminum foam is proposed as a pounding energy absorber for protection against earthquake hazards. The composite was created in two versions, with and without an adhesive interface. A series of experiments which reflect the conditions of seismic collision were performed: quasi-static compression, dynamic uniaxial compression and low-cycle dynamic compression with 10 loops of unloading at 10% strain. The composite material’s behavior was observed and compared with respect to uniform material specimens: polymer and metal foam. The experimental results showed that the maximum energy absorption efficiency in the case of the new material with the bonding layer was improved by 34% and 49% in quasi-static and dynamic conditions, respectively, in comparison to a sole polymer bumper. Furthermore, the newly proposed composites dissipated from 35% to 44% of the energy absorbed in the cyclic procedure, whereas the polymer specimen dissipated 25%. The capacity of the maintenance of the dissipative properties throughout the complete low-cycle loading was also satisfactory: it achieved an additional 100% to 300% of the energy dissipated in the first loading–unloading loop.

The article presents a novel application of the most up-to-date computational approach, i.e., artificial intelligence, to the problem of the compression of closed-cell aluminium. The objective of the research was to investigate whether the phenomenon can be described by neural networks and to determine the details of the network architecture so that the assumed criteria of accuracy, ability to prognose and repeatability would be complied. The methodology consisted of the following stages: experimental compression of foam specimens, choice of machine learning parameters, implementation of an algorithm for building different structures of artificial neural networks (ANNs), a two-step verification of the quality of built models and finally the choice of the most appropriate ones. The studied ANNs were two-layer feedforward networks with varying neuron numbers in the hidden layer. The following measures of evaluation were assumed: mean square error (MSE), sum of absolute errors (SAE) and mean absolute relative error (MARE). Obtained results show that networks trained with the assumed learning parameters which had 4 to 11 neurons in the hidden layer were appropriate for modelling and prognosing the compression of closed-cell aluminium in the assumed domains; however, they fulfilled accuracy and repeatability conditions differently. The network with six neurons in the hidden layer provided the best accuracy of prognosis at $MARE\le 2.7\%$ but little robustness. On the other hand, the structure with a complexity of 11 neurons gave a similar high-quality of prognosis at $MARE\le 3.0\%$ but with a much better robustness indication (80%). The results also allowed the determination of the minimum threshold of the accuracy of prognosis: $MARE\ge 1.66\%$. In conclusion, the research shows that the phenomenon of the compression of aluminium foam is able to be described by neural networks within the frames of made assumptions and allowed for the determination of detailed specifications of structure and learning parameters for building models with good-quality accuracy and robustness.

A novel double-emitter ultrasonic system for distance measurements based on the correlation method is presented. The proposed distance measurement method may be particularly useful in difficult conditions, e.g. for media parameters undergoing fast changes or in cases when obstacles and mechanical interference produce false reflections. The system is a development of a previously studied single-head idea. The present article covers a comparison of the two systems in terms of efficiency and precision. Experimental research described in this paper indicated that adding the second head improved the measurement exactness – standard deviation decreased by 40%. The correlation method is also described in detail, also giving the criterion for the quality of the measurement signal.

In this article, we present a proposition of a model of the compressive behaviour of open-cell aluminium with relation to the material apparent density. The research was based on experimental data from uniaxial compression tests conducted for two sample lots. These results were analysed with the use of neural networks in a specially designed algorithm. The main criterion for choosing a satisfactory approximation was mean absolute relative error MARE < 5%. As a result of the analysis, the sought relation was extracted and is presented as a proposition of a new ANN model of the compressive stress-strain relationship for aluminium sponge.