A-BiYOLOv9: An Attention-Guided YOLOv9 Model for Infrared-Based Wind Turbine Inspection

Maintenance scheduling is essential and crucial for wind turbines to avoid breakdowns and reduce maintenance costs. Many maintenance models have been developed for wind turbines' maintenance planning, such as corrective, preventive and predictive maintenance. Due to communities' dependence on wind turbines for electricity needs, preventative maintenance is the most widely used method for maintenance scheduling. The downside to using this approach is that preventive maintenance is often done in fixed intervals, which is inefficient. In this paper, a more detailed maintenance plan for a 2MW wind turbine has been developed. The paper's focus is to minimize a wind turbine's maintenance cost based on a wind turbine's reliability model. This study uses a two-layer optimization framework: Fibonacci and Genetic Algorithm (GA). The first layer in the optimization method (Fibonacci) finds the optimal number of preventive maintenance required for the system. In the second layer, the optimal times for preventative maintenance and optimal components to maintain have been determined to minimize maintenance costs. The Monte Carlo simulation estimates wind turbine component failure times using their lifetime distributions from the reliability model. The estimated failure times are then used to determine the overall corrective and preventive maintenance costs during the system's lifetime. Finally, an optimal preventive maintenance schedule is proposed for a 2MW wind turbine using the presented method. The method used in this paper can be expanded to a wind farm or similar engineering systems.

Jet engines are critical assets in aircraft, and their availability is crucial in the modern aircraft industry. Therefore, their maintenance scheduling is one of the major tasks an airline has to make during an engine’s lifetime. A proper engine maintenance schedule can significantly reduce maintenance costs without compromising the aircraft's reliability and safety. Different maintenance scheduling approaches have been used for jet engines, such as corrective, preventive, and predictive maintenance strategies. Regarding the safety demands in aircraft industries, preventive maintenance is a frequent maintenance method for jet engines. However, preventive maintenance schedules are often use fixed maintenance intervals, which is usually suboptimal. This paper focuses on minimizing a jet engine's overall maintenance cost by optimizing its preventive maintenance schedule based on an engine’s comprehensive reliability model. A hierarchical optimization framework including the golden section search and genetic algorithms is applied to find the optimal set of preventive maintenance number and their times and the components to be replaced at those times during the jet engine's overall lifetime. The Monte Carlo simulation is used to estimate the engine’s failure times using their lifetime distributions from the reliability model. The estimated failure times are then used to determine the engine's overall corrective and preventive maintenance costs during its lifetime. Finally, an optimal preventive maintenance schedule is proposed for an RB 211 jet engine using the presented method. In the end, comparing the proposed method's overall maintenance cost with two other maintenance methods demonstrates the proposed schedule's effectiveness. The method presented in this paper is generic, and it can be used for other similar engineering systems.

Real-world operation conditions and on-road emissions of Beijing diesel buses measured by using portable emission measurement system and electric low-pressure impactor

The study of reliability, availability and control of industrial manufacturing machines is a constant challenge in the industrial environment. This paper compares the results offered by several maintenance strategies for multi-stage industrial manufacturing machines by analysing a real case of a multi-stage thermoforming machine. Specifically, two strategies based on preventive maintenance, Preventive Programming Maintenance (PPM) and Improve Preventive Programming Maintenance (IPPM) are compared with two new strategies based on predictive maintenance, namely Algorithm Life Optimisation Programming (ALOP) and Digital Behaviour Twin (DBT). The condition of machine components can be assessed with the latter two proposals (ALOP and DBT) using sensors and algorithms, thus providing a warning value for early decision-making before unexpected faults occur. The study shows that the ALOP and DBT models detect unexpected failures early enough, while the PPM and IPPM strategies warn of scheduled component replacement at the end of their life cycle. The ALOP and DBT strategies algorithms can also be valid for managing the maintenance of other multi-stage industrial manufacturing machines. The authors consider that the combination of preventive and predictive maintenance strategies may be an ideal approach because operating conditions affect the mechanical, electrical, electronic and pneumatic components of multi-stage industrial manufacturing machines differently.

ABSTRACTOff-road vehicles used in construction and agricultural activities can contribute substantially to emissions of gaseous pollutants and can be a major source of submicrometer carbonaceous particles in many parts of the world. However, there have been relatively few efforts in quantifying the emission factors (EFs) and for estimating the potential emission reduction benefits using emission control technologies for these vehicles. This study characterized the black carbon (BC) component of particulate matter and NOx, CO, and CO2 EFs of selected diesel-powered off-road mobile sources in Mexico under real-world operating conditions using on-board portable emissions measurements systems (PEMS). The vehicles sampled included two backhoes, one tractor, a crane, an excavator, two front loaders, two bulldozers, an air compressor, and a power generator used in the construction and agricultural activities. For a selected number of these vehicles the emissions were further characterized with wall-flow diesel particle filters (DPFs) and partial-flow DPFs (p-DPFs) installed. Fuel-based EFs presented less variability than time-based emission rates, particularly for the BC. Average baseline EFs in working conditions for BC, NOx, and CO ranged from 0.04 to 5.7, from 12.6 to 81.8, and from 7.9 to 285.7 g/kg-fuel, respectively, and a high dependency by operation mode and by vehicle type was observed. Measurement-base frequency distributions of EFs by operation mode are proposed as an alternative method for characterizing the variability of off-road vehicles emissions under real-world conditions. Mass-based reductions for black carbon EFs were substantially large (above 99%) when DPFs were installed and the vehicles were idling, and the reductions were moderate (in the 20–60% range) for p-DPFs in working operating conditions. The observed high variability in measured EFs also indicates the need for detailed vehicle operation data for accurately estimating emissions from off-road vehicles in emissions inventories.Implications: Measurements of off-road vehicles used in construction and agricultural activities in Mexico using on-board portable emissions measurements systems (PEMS) showed that these vehicles can be major sources of black carbon and NOX. Emission factors varied significantly under real-world operating conditions, suggesting the need for detailed vehicle operation data for accurately estimating emissions inventories. Tests conducted in a selected number of sampled vehicles indicated that diesel particle filters (DPFs) are an effective technology for control of diesel particulate emissions and can provide potentially large emissions reduction in Mexico if widely implemented.

As the demand for wind energy increases, industry and policymakers have been pushing to place larger wind turbines in denser wind farms. Furthermore, there are higher expectations for reliability of turbines, which require a better understanding of the complex interaction between wind turbines and the fluid flow that drives them. As a test platform, we used the Whisper 500 residential scale wind turbine to support structural and atmospheric modeling efforts undertaken to improve understanding of these interactions. The wind turbine’s flexible components (blades, tower, etc.) were modeled using finite elements, and modal tests of these components were conducted to provide data for experimental validation of the computational models. Finally, experimental data were collected from the wind turbine under real-world operating conditions. The FAST (Fatigue, Aerodynamics, Structures, and Turbulence) software developed at the National Renewable Energy Laboratory was used to predict total system performance in terms of wind input to power output along with other experimentally measurable parameters such as blade tip and tower top accelerations. This paper summarizes the laboratory and field test experiments and concludes with a discussion of the models’ predictive capability. LA-UR-12-24832.KeywordsWind energyModal analysisFASTAirfoilDynamic testing

The present study provides firstly a comprehensive review of studies on measuring the impacts of different biodiesel blends on exhaust emissions characteristics of urban busses under real-world operating conditions. Secondly, this paper discusses the errors that can be made in conducting case studies. Thirdly and finally, it shows lessons learned and provides guidelines to setup case studies, conduct the measurements, perform the statistical analysis and report the results to policy makers and the wider audience. To achieve climate change mitigation targets, using alternative fuels, e.g., biodiesel, hydrogen or electricity for the urban fleets requires an in-depth analysis of the impacts under real-world operating conditions. Such experiments are generally very complex as numerous factors could directly or indirectly interfere with the results produced and potentially jeopardize the integrity of the research and the conclusions drawn. Results of the present research show that some vital parameters were ignored by many of the studies performed including the statistical uncertainties, driving cycle uncertainties and fuel uncertainties. Lack of appropriate experimental designs or clear assertions about the level of significance for differences in emissions/fuel consumption between alternative fuels (i.e. biodiesel) and the reference fuel used (i.e., diesel) could be regarded as the main weaknesses. Moreover, many other overarching and very influential factors (e.g., covariates/confounders) can interfere with the research outcomes as these were mostly overlooked by the reviewed studies. A careful and complete experimental design for assessments of alternative fueled vehicles are critical when conducting real-world operating condition tests. The study findings help to formulate the guidelines for assessing real-world operating condition experiments to achieve the most feasibly and meaningful research outcomes that will have significant implication for local and global policy makers. The guidelines are of use for all types of research studies that want to evaluate the effects of alternative fuels for any transportation fleet.

Gearbox vibration data collected under different operating conditions often suffer from inconsistent feature distribution and the challenge of removing noise components, which complicates fault diagnosis in real-world engineering applications. This paper proposes a deep transfer learning-based fault diagnosis method that addresses these challenges by leveraging discriminative feature extraction and improved domain adversarial neural networks. First, labelled and unlabelled vibration signals are organised into datasets using a fixed-length data segmentation method. To mitigate the interference of noise signals in real-world operating conditions, we employ a convolutional block attention module and a discriminant loss function-assisted feature extractor to extract distinguishing features. Next, to tackle the problem of inconsistent feature distribution, the multiple kernel maximum mean discrepancy is used to align the global distributions of the source and target domains, while an adversarial method is applied to align the sub-domain distributions between the two domains. Finally, experiments are conducted on an open-variable condition gearbox fault dataset. The results demonstrate that the proposed method achieves an average recognition accuracy of over 96.12%, confirming its effectiveness and superiority over other diagnostic methods. This algorithm offers significant value in practical engineering, providing a robust approach for fault diagnosis in industrial gearboxes with varying operating conditions and noise, thus improving maintenance efficiency and reducing downtime.

Majority of the recent approaches for text-independent speaker recognition apply attention or similar techniques for aggregation of frame-level feature descriptors generated by a deep neural network (DNN) front-end. In this paper, we propose methods of convolutional attention for independently modelling temporal and frequency information in a convolutional neural network (CNN) based front-end. Our system utilizes convolutional block attention modules (CBAMs) [1] appropriately modified to accommodate spectrogram inputs. The proposed CNN front-end fitted with the proposed convolutional attention modules outperform the no-attention and spatial-CBAM baselines by a significant margin on the VoxCeleb [2], [3] speaker verification benchmark. Our best model achieves an equal error rate of 2.031% on the VoxCeleb1 test set, which is a considerable improvement over comparable state of the art results. For a more thorough assessment of the effects of frequency and temporal attention in real-world conditions, we conduct ablation experiments by randomly dropping frequency bins and temporal frames from the input spectrograms, concluding that instead of modelling either of the entities, simultaneously modelling temporal and frequency attention translates to better real-world performance.

Making the correct maintenance strategy decision for industrial multistage machines (MSTM) is a constant challenge for industrial manufacturers. Preventive maintenance strategies are the most popular and provide interesting results but cannot prevent unexpected failures and consequences, such as time lost production (TLP). In these cases, a predictive maintenance strategy should be used to maintain the appropriate level of operation time. This research aims to present a model to identify the component that failed before its mean time to failure (MTTF) and, depending on whether the cause of the failure is known, propose the use of a predictive maintenance strategy and further decision-making to ensure the highest possible value from operating time. Also, it is necessary to check the reliable value of MTTF before taking certain decisions. For this research, a real case study of a MSTM was characterized component by component, setting the individual maintenance times. The initial maintenance strategy used for all the components is the preventive programming maintenance (PPM). If a component presents an unexpected failure, a method is proposed to decide whether the maintenance strategy should be changed, adding a predictive maintenance strategy to monitor said component. The research also provides a trust level to evaluate the reliable value of MTTF of each component. The authors consider this approach very useful for machine manufacturers and end users.

Approach for a Holistic Predictive Maintenance Strategy by Incorporating a Digital Twin

Wind turbine failure prediction and health assessment based on adaptive maximum mean discrepancy

PurposeThe purpose of this study is to optimize the operational efficiency of the entire system by developing a reasonable maintenance strategy for wind turbines that improves component reliability and safety while reducing maintenance costs.Design/methodology/approachA hybrid incomplete preventive maintenance (PM) model based on boundary intensity process is established to give dynamic PM intervals for wind turbines using an iterative method with reliability as a constraint; the selection method of PM and replacement is given based on the cost-effectiveness ratio, which in turn determines the optimal number of PM for wind turbines.FindingsThe reliability is used to obtain the components’ maintenance cycle, and the cost-effectiveness ratio is used to select the number of maintenance times, thus, getting the optimal maintenance strategy. The validity of this paper’s method is verified by arithmetic cases, which provides a new method for formulating a reasonable PM strategy for wind turbines.Practical implicationsThe wind turbine preventive maintenance strategy for Boundary intensity process proposed in this paper can scientifically formulate the maintenance strategy, optimize the cost-effectiveness per unit of time of the wind power generation system, and solve the problems of difficulty in formulating a reasonable maintenance strategy for the wind turbine components and high operation and maintenance costs.Originality/valueIn this paper, the authors describe the failure pattern by a Boundary intensity process, establish a hybrid incomplete PM model by introducing a failure intensity increment factor and an age reduction factor and establish a maintenance strategy optimization model with comprehensive consideration of reliability and cost-effectiveness ratio. Finally, the validity of the model in this paper is verified by arithmetic case analysis.Peer reviewThe peer review history for this article is available at: https://publons.com/publon/10.1108/ILT-05-2024-0153/

This paper presents a novel reference model designed to optimize the integration of preventive and predictive maintenance strategies for offshore wind farms (OWFs), enhancing operational decision-making. The model’s flexible and declarative architecture facilitates the incorporation of new constraints while maintaining computational efficiency, distinguishing it from existing methodologies. Unlike previous research that did not explore the intricate cost dynamics between predictive and preventive maintenance, our approach explicitly addresses the balance between maintenance expenses and wind turbine (WT) downtime costs. We quantify the impacts of these maintenance strategies on key operational metrics, including the Levelized Cost of Energy (LCOE). Using a constraint programming framework, the model enables rapid prototyping of alternative maintenance scenarios, incorporating real-time data on maintenance history, costs, and resource availability. This approach supports the scheduling of service logistics, including the optimization of vessel fleets and service teams. Simulations are used to evaluate the model’s effectiveness in real-world scenarios, such as handling the maintenance of up to 11 wind turbines per business day using no more than four service teams and four vessels, achieving a reduction in overall maintenance costs in simulated case of up to 32% compared to a solution that aims to prevent all downtime events. The prototype implementation as a task-oriented Decision Support System (DSS) further shows its potential in minimizing downtime and optimizing logistics, providing a robust tool for OWF operators.

Intermediate volatility organic compound (IVOC) emissions from a large cargo vessel were characterized under real-world operating conditions using an on-board measurement system. Test ship fuel-based emission factors (EFs) of total IVOCs were determined for two fuel types and seven operating conditions. The average total IVOC EF was 1003 ± 581 mg·kg-fuel-1, approximately 0.76 and 0.29 times the EFs of primary organic aerosol (POA) emissions from low-sulfur fuel (LSF, 0.38 wt % S) and high-sulfur fuel (HSF, 1.12 wt % S), respectively. The average total IVOC EF from LSF was 2.4 times that from HSF. The average IVOC EF under low engine load (15%) was 0.5-1.6 times higher than those under 36%-74% loads. An unresolved complex mixture (UCM) contributed 86.1 ± 1.9% of the total IVOC emissions. Ship secondary organic aerosol (SOA) production was estimated to be 546.5 ± 284.1 mg·kg-fuel-1; IVOCs contributed 98.9 ± 0.9% of the produced SOA on average. Fuel type was the dominant determinant of ship IVOC emissions, IVOC volatility distributions, and SOA production. The ship emitted more IVOC mass, produced higher proportions of volatile organic components, and produced more SOA mass when fueled with LSF than when fueled with HSF. When reducing ship POA emissions, more attention should be paid to commensurate control of ship SOA formation potential.