This paper presents a deep learning framework for analyzing on-board vibration response signals in infrastructure health monitoring. The proposed WaveletInception–BiGRU network uses a Learnable Wavelet Packet Transform (LWPT) for early spectral feature extraction, followed by one-dimensional Inception-Residual Network (1D Inception-ResNet) modules for multi-scale, high-level feature learning. Bidirectional Gated Recurrent Unit (BiGRU) modules then integrate temporal dependencies and incorporate operational conditions, such as the measurement speed. This approach enables effective analysis of vibration signals recorded at varying speeds, eliminating the need for explicit signal preprocessing. The sequential estimation head further leverages bidirectional temporal information to produce an accurate, localized assessment of infrastructure health. Ultimately, the framework generates high-resolution health profiles spatially mapped to the physical layout of the infrastructure. Case studies involving track stiffness regression and transition zone classification using real-world measurements demonstrate that the proposed framework significantly outperforms state-of-the-art methods, underscoring its potential for accurate, localized, and automated on-board infrastructure health monitoring. • WaveletInception-BiGRU proposed for on-board monitoring of railway infrastructure • WaveletInception extracts multi-scale local features from vibration signals • BiGRU captures temporal dependencies for localized health condition estimation • Late-stage fusion automates speed integration, eliminating manual feature engineering

Computer-aided simulations are routinely used to predict a prototype's performance. High-fidelity physics-based simulators might be computationally expensive for design and optimization, spurring the development of cheap deep-learning surrogates. The resulting surrogates often struggle to generalize and predict novel scenarios beyond their training domain. We propose a two-stage methodology addressing the challenge of generalization. It employs physics-based simulators, supplemented with ordinary differential equations integrated into the recurrent architecture, to learn the intrinsic dynamics. The proposed approach captures the inherent causality and generalizes the dynamics irrespective of a data source. The presented numerical experiments encompass five fundamental structural engineering scenarios, including beams on Winkler foundations based on Euler–Bernoulli and Timoshenko theories, beams under moving loads, and catenary-pantograph interactions in railways. The proposed methodology outperforms conventional recurrent methods and remains invariant to data sources, showcasing its efficacy. Numerical experiments highlight its prospects for design optimization, predictive maintenance, and enhancing safety measures.

This article develops and tests a self-contained railway track monitoring system that fits in existing vehicles without the need for speed and load control. Combining a train-borne laser Doppler vibrometer and axle box accelerometers enables synchronized measurements of train-track response under operational conditions. Utilizing a GPS antenna and video camera, we propose the multisignal processing method to obtain train-track vibrations with train position and speed. Then, we fuse the multiple signals to extract an impact index and a resonance index and further propose an interpretable anomaly detection strategy. We test the system on an operational line at 20–60 km/h under different working conditions and verify the detection results using information from conventional technologies. The impact index peaks near joints and welds, and the resonance index yields a good correlation with the measured track geometry. The developed solution achieves the detection, localization, and quantification of surface and support anomalies in railway tracks.

A primary challenge of physics-informed machine learning (PIML) is its generalization beyond the training domain, especially when dealing with complex physical problems represented by partial differential equations (PDEs). This paper aims to enhance the generalization capabilities of PIML, facilitating practical, real-world applications where accurate predictions in unexplored regions are crucial. We leverage the inherent causality and temporal sequential characteristics of PDE solutions to fuse PIML models with recurrent neural architectures based on systems of ordinary differential equations, referred to as neural oscillators. Through effectively capturing long-time dependencies and mitigating the exploding and vanishing gradient problem, neural oscillators foster improved generalization in PIML tasks. Extensive experimentation involving time-dependent nonlinear PDEs and biharmonic beam equations demonstrates the efficacy of the proposed approach. Incorporating neural oscillators outperforms existing state-of-the-art methods on benchmark problems across various metrics. Consequently, the proposed method improves the generalization capabilities of PIML, providing accurate solutions for extrapolation and prediction beyond the training data.

This paper proposes a novel framework for simulating the dynamics of beams on elastic foundations. Specifically, partial differential equations modeling Euler–Bernoulli and Timoshenko beams on the Winkler foundation are simulated using a causal physics-informed neural network (PINN) coupled with transfer learning. Conventional PINNs encounter challenges in handling large space–time domains, even for problems with closed-form analytical solutions. A causality-respecting PINN loss function is employed to overcome this limitation, effectively capturing the underlying physics. However, it is observed that the causality-respecting PINN lacks generalizability. We propose using solutions to similar problems instead of training from scratch by employing transfer learning while adhering to causality to accelerate convergence and ensure accurate results across diverse scenarios. The primary contribution of this paper lies in introducing a causality-respecting PINN loss function in the context of structural engineering and coupling it with transfer learning to enhance the generalizability of PINNs in simulating the dynamics of beams on elastic foundations. Numerical experiments on the Euler–Bernoulli beam highlight the efficacy of the proposed approach for various initial conditions, including those with noise in the initial data. Furthermore, the potential of the proposed method is demonstrated for the Timoshenko beam in an extended spatial and temporal domain. Several comparisons suggest that the proposed method accurately captures the inherent dynamics, outperforming the state-of-the-art physics-informed methods under standard L2-norm metric and accelerating convergence.

Microgrids (MGs) are sustainable solutions for rural zone electrification that use local renewable resources. However, only careful planning at the start of an MG project can ensure its future optimal operation. In this paper, a novel methodology for MG planning by using the uncertainty characterization of renewable resources and demand is presented. Additionally, a model of electricity consumption is proposed and applied in an isolated rural community. In such communities, consumption patterns typically need to be derived as model inputs because consumption measurements are not available for the planning stage. To obtain these inputs, clustering algorithms based on self-organizing maps (SOMs) and fuzzy c-means are used to classify the families of the community given sociodemographic information obtained via surveys. Subsequently, Markov chains (MCs) are employed to generate consumption patterns based on consumption measurements in some dwellings and surveys applied to the community. The nonlinearities and uncertainties associated with renewable resources and consumption are modeled by using prediction interval (PI) models. These PI models provide the required consumption and generation scenarios for deriving the optimal sizing and topological information to address the MG planning problem. The results of the robust planning approach based on scenarios are useful at the feasibility and design phases of an MG project. The proposed methodology is successfully applied to MG planning for a rural Mapuche community, where a conservative criterion was considered to minimize the investment risk. This criterion corresponds to the worst-case scenario in which the demand increases by 19.9% compared to that of the baseline scenario and a lower energy cost is obtained. However, the net present cost and operational costs increase by 14% and 11.75% compared to those of the baseline scenario, respectively.

This article proposes a new framework using physics-informed neural networks (PINNs) to simulate complex structural systems that consist of single and double beams based on Euler-Bernoulli and Timoshenko theories, where the double beams are connected with a Winkler foundation. In particular, forward and inverse problems for the Euler-Bernoulli and Timoshenko partial differential equations (PDEs) are solved using nondimensional equations with the physics-informed loss function. Higher order complex beam PDEs are efficiently solved for forward problems to compute the transverse displacements and cross-sectional rotations with less than 1e-3 % error. Furthermore, inverse problems are robustly solved to determine the unknown dimensionless model parameters and applied force in the entire space-time domain, even in the case of noisy data. The results suggest that PINNs are a promising strategy for solving problems in engineering structures and machines involving beam systems.

In this paper, a fuzzy interval-based method is proposed for solving the problem of rail defect detection relying on an on-board measurement system and a multiple spiking neural network architecture. Instead of outputting binary values (defect or not defect), all data will belong to both classes with different spreads that are given by two fuzzy intervals. The multiple spiking neural networks are used to capture different sources of uncertainties. In this paper, we consider uncertainties in the parameters of spiking neural networks during the training phase. The proposed method comprises two steps. In the first step, multiple sets of the firing times for both classes are obtained from multiple spiking neural networks. In the second step, the obtained multiple sets of firing times are fuzzy numbers and they are used to construct fuzzy intervals. The proposed method is showcased with the problem of rail defect detection. The numerical analysis indicates that the fuzzy intervals are suitable to make use of the information provided by the multiple spike neural networks. Finally, with the proposed method, we improve the interpretability of the decision making regarding the detection of anomalies.

After a natural disaster, a quick inspection of all damaged components is crucial to recover the functionality of distribution networks. Unmanned aerial vehicles (UAVs) can perform inspection tasks, particularly for damages that are difficult to access for human repair crews. Additionally, UAVs can monitor the transmission lines to find potential dangers and early-stage damages, and to monitor the road infrastructure to provide real-time information about traffic conditions so that repair crews can select the best ways to reach damages. Besides, due to unpredictable events during restoration, the UAV routing strategy (UAVRS) needs to be updated in real time. Thus, the proposed UAVRS in this article determines the optimal routes for the UAVs allocated to inspect damages as well as the optimal routes for the UAVs to monitor transmission lines and roads in real time for distribution networks. To tackle the multi-time-scale characteristic of the proposed UAVRS, a two-layer decision-making architecture is proposed. A bilevel programming problem is solved in the first layer for the large-time-scale problem, and a mixed-integer linear programming problem is solved for the small-time-scale problem in the second layer. A case study based on the distribution network in Zaltbommel and its neighbor areas, in The Netherlands, illustrates the effectiveness of our real-time method compared to the offline methods. Furthermore, different solvers are studied and compared in view of the real-time requirement.

The catenary insulator maintains electrical insulation between catenary and ground. Its defects may happen due to the long-term impact from vehicle and environment. At present, the research of defect detection for catenary insulator faces several challenges. 1) Localization accuracy is low, which causes the localized object to be incomplete or/and merge with unnecessary background. 2) Horizontal localization brings inevitable unnecessary information because horizontal box cannot fit well with the shape of insulator. 3) Supervised learning models for defects recognition are unreliable as the available defect samples are insufficient to train models well. To address these issues, this article proposes a novel two-stage defect detection method. In the localization stage, a novel localization network called TOL-Framework is constructed to reduce the background and realize tighter oriented localization. Compared with general basic framework Faster R-CNN, the TOL-Framework cascades a regression module inside basic framework and adds an external postprocess network, which is adversarially trained by standard insulators to refine the localization. These two novel steps greatly improve the oriented localization accuracy. In the defect detection stage, an adversarial reconstruction model that is trained only using normal samples is proposed to evaluate the defect states. A comparison with other methods is conducted using a dataset collected from a 60km section of the Changsha-Zhuzhou railway line in China. The results show the proposed method has the highest localization accuracy, and is effective for insulator defect detection.

The growing variety of data from condition monitoring of high-speed railways offer unprecedented opportunities to improve railway infrastructure maintenance. For condition monitoring of railway catenaries, this paper proposes a data-driven approach that uses a Bayesian network (BN) to integrate the inspection data from catenaries into a key performance indicator (KPI). The BN topology is structured based on the physical relationships among data types, including train speed, dynamic stagger and height of the contact wire, pantograph head acceleration, and pantograph-catenary contact force. The tailored performance indicators are individually defined and extracted from the five types of data as the BN input. As the output of the BN, the KPI is defined as the overall condition level of the catenary considering all defects that can be reflected by the data types. Finally, using historical inspection data and maintenance records from a section of the Beijing-Guangzhou high-speed line in China, the BN parameters are estimated to establish a probabilistic relationship between the input and output. An approach that applies the estimated BN to catenary condition monitoring is proposed. Testing of the BN-based approach using new inspection data shows that the output KPI can adequately represent the catenary condition, leading to a nearly 66.2% reduction in the false alarm rate of defect detection compared with current practice. It is also tested that when the input data quality is not ideal, the approach can still work acceptably on noisy data with a signal-to-noise ratio greater than 3 dB or with one type of data missing.

This paper presents a Pareto-based maintenance decision system for rail welds in a regional railway network. Weld health condition data are collected using a train in operation. A Hilbert spectrum-based approach is used for data processing to detect and assess the weld quality based on multiple registered dynamic responses in the axle box acceleration measurements. The assessment of the welds is stochastic in nature and variant over time, so a set of robust and predictive key performance indicators is defined to capture the weld degradation dynamics during a given maintenance period. Using a scenario-based approach, two objective functions are defined, performance and the number of weld replacements. Evolutionary multiobjective optimization is employed to optimize the objective functions so that the tradeoffs between performance and cost support decision-making for railway network maintenance. The results of the proposed methodology show that the infrastructure manager can localize field inspections and maintenance efforts on the area with the most critical welds. To showcase the capability of the proposed methodology, measurements from a regional railway network in Transylvania and Romania are employed.

In this paper, we propose a methodology based on signal processing and evolutionary multiobjective optimization to facilitate the maintenance decision making of infra-managers in regional railways. Using a train in operation (with passengers onboard), we capture the condition of the rails using Axle Box Acceleration measurements. Then, using Hilbert-Huang Transform, the locations where the major risks are detected and assessed with a degradation model. Finally, evolutionary multiobjective optimization is employed to solve the maintenance decision problem, and to facilitate the visualization of the trade-offs between number of interventions and performance. Real-life measurements from the track from Braşov to Zărneşti in Romania are included to show the methodology.

Microgrids are sustainable solutions for electrification of rural zones that can make use of their local renewable resources. In this paper, we propose a new method for microgrid planning which includes the effect of the uncertainties of the renewable resources explicitly. Fuzzy interval models are used because they can capture nonlinearities and systematically represent the uncertainties associated with renewable resources at a certain confidence level. Relying on interval fuzzy models and by considering a set of possible scenarios for the renewable resources, the solution to the microgrid planning problem is given through the optimal sizing and topology of the microgrid. This information, particularly the optimal sizes of generators and the economic analysis, is useful for the design phase of a microgrid project. The proposed methodology is applied to the microgrid planning of the rural Mapuche community, José Painecura, in Chile.

In this paper, we propose a deep convolutional neural network solution to the analysis of image data for the detection of rail surface defects. The images are obtained from many hours of automated video recordings. This huge amount of data makes it impossible to manually inspect the images and detect rail surface defects. Therefore, automated detection of rail defects can help to save time and costs, and to ensure rail transportation safety. However, one major challenge is that the extraction of suitable features for detection of rail surface defects is a non-trivial and difficult task. Therefore, we propose to use convolutional neural networks as a viable technique for feature learning. Deep convolutional neural networks have recently been applied to a number of similar domains with success. We compare the results of different network architectures characterized by different sizes and activation functions. In this way, we explore the efficiency of the proposed deep convolutional neural network for detection and classification. The experimental results are promising and demonstrate the capability of the proposed approach.

This paper discusses the applicability of Big Data techniques to facilitate maintenance decisions regarding railway tracks. Currently, in different countries, a huge amount of railway track condition-monitoring data is being collected from different sources. However, the data are not yet fully used because of the lack of suitable techniques to extract the relevant events and crucial historical information. Thus, valuable information is hidden behind a huge amount of terabytes from different sensors. In this paper, the conditions of the 5V's of Big Data (Volume, Velocity, Variety, Veracity and Value) in railway monitoring systems are discussed. Then, general methods that can be applied to facilitate the decision of efficient railway track maintenance are proposed for railway track condition monitoring. As a benchmark, axle box acceleration (ABA) measurements in the Dutch tracks are used, and generic reduction formulations to address new relevant information and handle failures are proposed.

This paper presents an automatic method for detecting railway surface defects called “squats” using axle box acceleration (ABA) measurements on trains. The method is based on a series of research results from our group in the field of railway engineering that includes numerical simulations, the design of the ABA prototype, real-life implementation, and extensive field tests. We enhance the ABA signal by identifying the characteristic squat frequencies, using improved instrumentation for making measurements, and using advanced signal processing. The automatic detection algorithm for squats is based on wavelet spectrum analysis and determines the squat locations. The method was validated on the Groningen-Assen track in The Netherlands and accurately detected moderate and severe squats with a hit rate of 100%, with no false alarms. The methodology is also sensitive to small rail surface defects and enables the detection of squats at their earliest stage. The hit rate for small rail surface defects was 78%.

This paper presents a model comparison of a fixed speed wind turbine (FSWT) operating on a real wind farm. By relying on real data obtained from a wind farm operating in the Chilean Interconnected System, three different models are identified and analyzed. First, a phenomenological model based on physical principles governing the production of electricity from wind power is considered. This model is fine-tuned in accordance with practical considerations, such as wind correction factors. Then, a linear model and a Takagi & Sugeno (T&S) fuzzy model are identified. From the experimental results, the linear model is the simplest one, but also the one that presents the worst performance indexes. The best prediction capability is obtained with the T&S model; however, in terms of interpretability, the phenomenological model outperforms the other two black-box models.

In this paper, two methods for generating the daily load profile and forecasting in isolated small communities are proposed. In these communities, the energy supply is difficult to predict because it is not always available, is limited according to some schedules and is highly dependent on the consumption behavior of each community member. The first method is proposed to be used before the implementation of the microgrid in the design state, and it includes a household classifier based on a Self Organizing Map (SOM) that provides load patterns by the use of the socio-economic characteristics of the community obtained in a survey. The second method is used after the implementation of the microgrid, in the operation state, and consists of a neural network with on-line learning for the load forecasting. The neural network model is trained with real-data of load and it is designed to stay adapted according to the availability of measured data. Both proposals are tested in a real-life microgrid located in Huatacondo, in northern Chile (project ESUSCON). The results show that the estimated daily load profile of the community can be very well approximated with the SOM classifier. On the other hand, the neural network can forecast the load of the community reasonably well two-days ahead. Both proposals are currently being used in a key module of the energy management system (EMS) in the real microgrid to optimize the real uninterrupted load for 24-hour energy supply service.

In this paper we compare four methods for decentralized Kalman filtering for distributed-parameter systems, which after spatial and temporal discretization, result in large-scale linear discrete-time systems. These methods are: parallel information filter, distributed information filter, distributed Kalman filter with consensus filter, and distributed Kalman filter with weighted averaging. These filters are suitable for sensor networks, where the sensor nodes perform not only sensing and computations, but also communicate estimates among each other. We consider an application of sensor networks to a heat conduction process. The performance of the decentralized filters is evaluated and compared to the centralized Kalman filter.

In this paper, a dynamic vehicle routing problem (DVRP) is solved based on hybrid predictive control strategy with an objective function that includes two dimensions: user and operator costs. To handle some undesired assignments for the users, a new objective function is designed, able to carry out the fact that some users can become particularly annoyed if their service is postponed. Genetic algorithms are proposed for efficiently solving the DVRP. Fuzzy clustering is applied for computing trip patterns from historical data under more realistic scenarios. An illustrative experiment through simulation of the process is presented to show the potential benefits (mainly for users) of the new design.

In the paper, the hybrid predictive control based on a fuzzy model is presented. The identification methodology for a nonlinear system with discrete state-space variables by combining fuzzy clustering and principal component analysis is proposed. The fuzzy model is used for hybrid predictive control design where the optimization problem is solved by the use of genetic algorithms. An illustrative experiment on a hybrid tank system is conducted to present the benefits of the proposed approach.