Most rolling element bearing (REB) fault diagnosis algorithms are evaluated on the Case Western Reserve University (CWRU) dataset for its popularity and simplicity. However, accuracy CWRU is overly saturated; it nearly up to 100%. In this study, an input feature mappings (IFMs)-based deep residual network (ResNet) proposed conduct detailed comprehensive REB with complicated dataset. Firstly, a new data preprocessing method named as signal-to-IFMs automatically extract features from raw...

Deep learning-based helmet detection is a common aspect of industrial safety production. The existing hard hat data set covers large number open-air scenes, and lacks low-light scenes such as underground mines tunnel projects. Therefore, it cannot meet the requirements under low light. To develop dataset better suited for this purpose, we first collected multiple images in harsh light to supplement data. Then, analyzed compared network structure model features current phase-based detectors,...

Accurate prediction of fuel consumption is critical for achieving efficient and low-carbon ship operations. However, the variability marine environment introduces significant challenges, as it leads to dynamic changes in monitoring data, complicating real-time precise prediction. To address this issue, authors proposed an incremental learning-based framework enhance adaptability temporal dependencies data. The dynamically adjusts a dual adaption mechanism input features target labels while...

Simulation technology has been used to train marine engineers for decades. However, traditional marine-engine training systems (METSs) possess low immersion levels that seriously limit their effect. To overcome this problem, we developed a head-mounted, display-based immersive virtual reality (IVR) METS based on HTC Vive Pro hardware. Like the METS, it and automatic evaluation functions. Moreover, can reproduce, with high degrees of immersion, all details real ship engine room vividly...

Machine learning is an effective way to realise the condition monitoring of marine machinery. However, it challenging this purpose based on supervised in practice due lack labelled data. To overcome problem, we propose use isolation forest decay detection a gas turbine with normal Besides, consider impact data contamination for first time compared previous literatures. We also experiment same datasets support vector description (SVDD) as comparison. The results show that very suitable...

The semi-physical marine engine room simulator is one of the necessary equipment for training engineers worldwide. However, its maintenance has become a problem long time due to complexity and vulnerability. Although fails frequently, failures are usually small-scale communication failures, which easy solve experts. For such kind micro faults, cost field by experts very high journey. Generally, most manufactures will choose assist users' telephone or email firstly. If it does not work, they...

This paper researches the heat transfer equation and thermal balance of a shell-and-tube evaporator; constructs an accurate mathematical model for derives equations including detailed calculation methods all coefficients, such as refrigerant side coefficient, water kinematic viscosity, density, specific enthalpy. Adopting this approach involves fitting relationships between conductivity, enthalpy R134a refrigerants in saturated vapor liquid states. The superheated gas enthalpy, temperature...

Abstract In this article, a multilabel support vector machine (SVM)-based approach is investigated to address the simultaneous decay detection of marine propulsion system. To verify performance algorithm, we perform some experiments using simulation dataset from real-data validated numerical simulator Frigate. particular, try train model without data, considering great difficulty obtaining data in practice. The experimental results show that proposed can identify complex modes system...

To solve marine seawater cooling system's faults better, the fault pattern recognition model of system is established. Firstly, structure and typical are analyzed, modes divided. Then LMS learning rules selected as algorithm SOM neural network, sample set constructed by using relevant state parameters collected from real ship to train network. The training results show that has satisfactory clustering effect. Finally, identification verified test data, can accurately diagnose mode system.

The semi physical marine engine room simulator is an important equipment widely used in the training of engineers worldwide. However, it only simulates a part control consoles room, which not conducive to trainees' cognition actual environment. Considering this problem, we develop immersive virtual based on reality technology as complement traditional simulator. 3D built according mother ship, so can match perfectly. combination them beneficial for trainees understand real environment while...

Deployment of fast response craft at sea is essential for a range littoral missions including maritime security and search-and-rescue. Traditional launch recovery like rigid-hull boats (RHB) employ side davits cranes. In light evolving technology, multi-role navy ships the Republic Singapore Navy’s Independence-class Littoral Mission Vessel have incorporated stern ramp Launch Recovery Systems (LARS) to overcome slower deployment times associated with cranes davits. Twin LARS are suitable...

Summary For carbonate reservoirs in the Tarim Basin of China, faults, fractures, caves and their combinations are main reservoir space. Faults accompanied rifts effectively rebuild reservoirs. Therefore, effective fault detection is very important for exploration development reservoir. Traditional geometric attributes sensitive to structural information such as faults strata, well sequence features riverway. These messages often mixed together, making it difficult distinguish separately....