This paper proposes a computationally effective framework for load‐dependent optimal sensor placement in large‐scale civil engineering structures subjected to moving loads. Two common problems are addressed: selection of modes to be monitored and computational effectiveness. Typical sensor placement methods assume that the set of modes to be monitored is known. In practice, determination of such modes of interest is not straightforward. A practical approach is proposed that facilitates the selection of modes in a quasi‐automatic way based on the structural response at the candidate sensor locations to typical operational loads. The criterion used to assess sensor placement is based on Kammer's Effective Independence (EFI). However, in contrast to typical implementations of EFI, which treat the problem as a computationally demanding discrete problem and use greedy optimization, an approach based on convex relaxation is proposed. A notion of sensor density is applied, which converts the original combinatorial problem into a computationally tractable continuous optimization problem. The proposed framework is tested in application to a real tied‐arch railway bridge located in central Poland.

The area affected by the earthquake is vast and often difficult to entirely cover, and the earthquake itself is a sudden event that causes multiple defects simultaneously, that cannot be effectively traced using traditional, manual methods. This article presents an innovative approach to the problem of detecting damage after sudden events by using an interconnected set of deep machine learning models organized in a single pipeline and allowing for easy modification and swapping models seamlessly. Models in the pipeline were trained with a synthetic dataset and were adapted to be further evaluated and used with unmanned aerial vehicles (UAVs) in real-world conditions. Thanks to the methods presented in the article, it is possible to obtain high accuracy in detecting buildings defects, segmenting constructions into their components and estimating their technical condition based on a single drone flight.

Nowadays, consumer electronics offer computer-vision-based (CV) measurements of dynamic displacements with some trade-offs between sampling frequency, resolution and low cost of the device. This study considers a consumer-grade smartphone camera based on complementary metal-oxide semiconductor (CMOS) technology and investigates the influence of its hardware limitations on the estimation of dynamic displacements, modal parameters and stiffness parameters of bolted connections in a laboratory structure. An algorithm that maximizes the zero-normalized cross-correlation function is employed to extract the dynamic displacements. The modal parameters are identified with the stochastic subspace identification method. The stiffness parameters are identified using a model-updating technique based on modal sensitivities. The results are compared with the corresponding data obtained with accelerometers and a laser distance sensor. The CV measurement allows lower-order vibration modes to be identified with a systematic (bias) error that is nearly proportional to the vibration frequency: from 2% for the first mode (9.4 Hz) to 10% for the third mode (71.4 Hz). However, the measurement errors introduced by the smartphone camera have a significantly lower influence on the values of the identified stiffness parameters than the numbers of modes and parameters taken into account. This is due to the bias–variance trade-off. The results show that consumer-grade electronics can be used as a low-cost and easy-to-use measurement tool if lower-order modes are required.

Visual inspection of civil infrastructure for structural health assessment, as performed by structural engineers, is expensive and time-consuming. Therefore, automating this process is highly attractive, which has received significant attention in recent years. With the increasing capabilities of computers, deep neural networks have become a standard tool and can be used for structural health inspections. A key challenge, however, is the availability of reliable datasets. In this work, the U-net and DeepLab v3+ convolutional neural networks are trained on a synthetic Tokaido dataset. This dataset comprises images representative of data acquired by unmanned aerial vehicle (UAV) imagery and corresponding ground truth data. The data includes semantic segmentation masks for both categorizing structural elements (slabs, beams, and columns) and assessing structural damage (concrete spalling or exposed rebars). Data augmentation, including both image quality degradation (e.g., brightness modification, added noise) and image transformations (e.g., image flipping), is applied to the synthetic dataset. The selected neural network architectures achieve excellent performance, reaching values of 97% for accuracy and 87% for Mean Intersection over Union (mIoU) on the validation data. It also demonstrates promising results in the semantic segmentation of real-world structures captured in photographs, despite being trained solely on synthetic data. Additionally, based on the obtained results of semantic segmentation, it can be concluded that DeepLabV3+ outperforms U-net in structural component identification. However, this is not the case in the damage identification task.

The article presents a water-cooling system for photovoltaic (PV) modules using a two-axis tracking system that tracks the apparent position of the Sun on the celestial sphere. The cooling system consists of 150 adjustable spray nozzles that cool the bottom layer of PV modules. The refrigerant is water taken from a tank with a capacity of 7 m 3. A water recovery system reduces its consumption with efficiency of approximately 90%. The experimental setup consists of a full-size photovoltaic installation made of 10 modules with an output power of 3.5 kWp combined with a tracking system. The article presents an analysis of the cooling system efficiency in various meteorological conditions. Measurements of energy production were performed in the annual cycle using three different types of photovoltaic installations: stationary, two-axis tracking system and two-axis tracking system combined with the cooling system.