Engine Fault Research Articles

Fault Detection of Diesel Engines Using Artificial Neural Network

<div>This study presents a method for identifying the reliability state of diesel engines by utilizing artificial neural networks (ANNs). The Sulzer 6AL20/24 marine diesel engine was selected as the test subject for this research. Vibration signals were collected during tests conducted on a laboratory test stand under normal operating conditions and during simulations of six different engine faults. Next, the recorded signals were analyzed and transformed into labeled samples for supervised learning. In this phase, the time histories of the vibration signals were divided into segments and augmented, with several key features calculated for each segment. Highly correlated signals were excluded from further analysis based on the Pearson correlation coefficient. The processed samples were then used to train and fine-tune the ANN. The trained ANN was subsequently used to identify the engine’s reliability state and classify the present fault type. To evaluate the effectiveness of the proposed method, the results obtained from the ANN were compared with those from technical state space identification and other machine learning classifiers. Finally, the ANN was tested on the reliability state of the engine in scenarios where the simulated fault was not present in the training dataset. A detailed discussion and analysis of the proposed identification method’s performance are presented in this article’s concluding sections.</div>

Knowledge-Based Intelligent System for Diagnosing Three-Wheeled Motorcycle Engine Faults

Three-wheeled motor engine damage is one of the most serious problems with all motorcycles. When problems appear, it becomes difficult for users to repair and diagnose faults because knowledge about machine breakdown symptoms is minimal. Most motorcycle repair shops don’t have mechanics who understand tricycle motorbike engines, so they are less accurate in diagnosing damage symptoms, only based on estimates. Three-wheeled motorbikes have several differences in structure and spare parts compared to motorcycles because tricycle motorbikes have an axle like a car. For this problem, an information system is needed with a method that combines an expert's experience, expertise, and knowledge to develop expert system applications based on several cases that have been experienced and are known as case-based reasoning. This research aims to produce a web-based expert system to diagnose and solve tricycle motorbike engine damage problems. The case-based reasoning method with the K-Nearest Neighbor algorithm is used to assist in analyzing engine damage and give solutions to the issues in three-wheeled motorbike engines. Using two methods is appropriate because of the answers found and the similarities calculated by the cosine similarity method, which experts then review to get the proper solution. From testing using 20 samples of diagnostic data, an accuracy percentage of 85% was obtained. The calculation result for precision is 85%, and recall is 85%.

Research on Fault Diagnosis Method for Marine Diesel Engines Based on Multi-Scale Attention Mechanism Transformer

In modern intelligent shipping, ensuring the stable and reliable technical condition of marine diesel engines is critical for safe and efficient vessel operations. Conventional fault diagnosis approaches and many existing Transformer-based methods often focus on single-scale features, potentially overlooking subtle fault indicators and reducing diagnostic accuracy under complex working conditions. To address these limitations, this paper proposes a Multi-Scale Attention Transformer (MSAT) model that integrates both high- and low-resolution attention mechanisms. This multi-scale strategy enhances the extraction of detailed and coarse-grained features, improving the model’s capacity to detect and characterize complex diesel engine faults. Additionally, an optimized Nadam optimizer is employed to refine convergence speed and accuracy, surpassing the Adam-based baseline by 0.71%. Rigorous testing on a publicly available diesel engine fault dataset demonstrates that the MSAT model achieves a diagnostic accuracy of 99.86% at a 60 dB signal-to-noise ratio (SNR), outperforming established models such as GRU and LSTM by more than 1%. Even under severe noise interference (0 dB SNR), the model maintains a high accuracy of 96.86%, highlighting its robustness and suitability for real-time monitoring in challenging marine environments. By quantitatively validating these improvements in diagnostic accuracy and noise resistance, this work offers a novel and effective solution for predictive maintenance and operational condition assessment of marine diesel engines, contributing to the reliability and safety of intelligent shipping systems.

Early warning of vibration faults of marine diesel engine based on improved VMD

ABSTRACT To improve the prediction accuracy of vibration signals from marine diesel engines, a novel method is proposed. This method utilises the rime optimisation algorithm (RIME) to select the variable modal decomposition (VMD) parameters for data reconstruction in order to optimise the bidirectional gated recirculation unit (BIGRU). The RIME-VMD-BIGRU model employs sample entropy as the fitness function. RIME determines the optimal k and α values for VMD, and the vibration signals are decomposed into subsequences. After noise reduction, these subsequences are reconstructed. A BIGRU prediction model is then built for the reconstructed components. The prediction results of these components are combined to provide vibration prediction. Case studies show that the proposed model achieves smaller prediction errors and higher accuracy compared to other models. It effectively provides early warnings for diesel engine vibration signals.

Non-Invasive Techniques for Monitoring and Fault Detection in Internal Combustion Engines: A Systematic Review

This article provides a detailed analysis of non-invasive techniques for the prediction and diagnosis of faults in internal combustion engines, focusing on the application of the Proknow-C and Methodi Ordinatio systematic review methods. Initially, the relevance of these techniques in promoting energy sustainability and mitigating greenhouse gas emissions is discussed, aligning with the Sustainable Development Goals (SDGs) of Agenda 2030 and the Paris Agreement. The systematic review conducted in the subsequent sections offers a comprehensive mapping of the state of the art, highlighting the effectiveness of combining these methods in categorizing and systematizing relevant scientific literature. The results reveal significant advancements in the use of artificial intelligence (AI) and digital signal processors (DSP) to improve fault diagnosis, in addition to highlighting the crucial role of non-invasive techniques such as the digital twin in minimizing interference in monitored systems. Finally, concluding remarks point towards future research directions, emphasizing the need to develop the integration of AI algorithms with digital twins for internal combustion engines and identify gaps for further improvements in fault diagnosis and prediction techniques.

Data-driven model for marine engine fault diagnosis using in-cylinder pressure signals

Effective diagnosis of marine engines that is crucial for safe and reliable ship operations requires fidel tools for the identification of critical faults. However, the unavailability of extensive measured data-sets corresponding to engine faulty conditions renders the development of such tools challenging. This study aims to develop a data-driven fault estimation model considering information extraction methods and regression techniques of low computational effort, namely multiple linear and polynomial regression. The required data-sets for healthy and faulty conditions for four critical faults and their combinations are generated by employing a calibrated zero-dimensional thermodynamic model representing a marine four stroke medium speed diesel engine, which is validated against engine shop trial measurements. Fourier analysis of the derived in-cylinder pressure profiles is employed to calculate the coefficients of the harmonic orders. Several harmonics coefficients sets are used as input to the regression models to estimate the severity of the four considered faults. The results demonstrate that initial 20 harmonics are sufficient to effectively estimate the severity for each fault, whereas polynomial regression is highly effective, exhibiting R 2 values greater than 98 % . This study provides insights on the data-driven simultaneous faults severity estimation, and as such it impacts the advancement of cost-effective diagnostic methods for marine engines.

Fault-tolerant control of nonlinear cluster system for fixed-wing UAV piston engine faults based on hierarchical architecture

Due to the information exchange between fixed-wing UAV clusters, the fault and disturbance information generated when the actuator, including the engine, is transmitted from the faulty UAVs to the healthy UAVs. Given this situation, this paper proposes a cluster fault-tolerant control strategy based on hierarchical architecture. Firstly, this paper analyzes the influence of vibration caused by engine failure on fixed-wing UAVs and illustrates the necessity of considering this factor with simulation experiments. Secondly, this paper designs a hierarchical architecture, which divides the distributed control of the fixed-wing cluster system into consistency control of the upper layer and fault-tolerant tracking control of the lower layer, so that the fault information will only exist in the faulty UAVs, ensuring that the healthy UAVs will not be affected, thus avoiding the occurrence of fault propagation and improving the overall fault tolerance of the cluster. Finally, by analyzing the stability of the Lyapunov function, it is shown that this method can still follow the leader in systems with the followers' failures, and the effectiveness of this method is verified by comparing simulation results.

Recognition and location of rotor–stator rub-impact failure of aircraft engine with periodicity feature of vibration signals

To precisely recognize and locate a rub-impact fault of aircraft engine, an approach combining Hilbert envelope (HE), adaptive motion window (AMW), and margin factor (MF) (HE-AMW-MF) was presented. HE can reflect the change rule and tendency of vibration signals and simplify the complexity of calculation of fault diagnosis. Therefore, the paper starts with the calculation of HE of signal and takes it in place of original signal for subsequent analysis. As vibration signal caused by rub-impact fault is often accompanied by periodic impact, the paper takes advantage of the characteristic of periodic impact (collision period instead of rotation period) to adaptively determine a length and step size of motion window and exactly distills the rub-impact fault characteristics of each period. Due to the sensitivity of margin factor (MF) to the trait of wearing and impact, it is brought in to represent the fault feature information contained in signal in each motion window. Equally, mean value of all MFs is calculated and built into feature vector to reduce an interference of noise. Finally, the feature vector is combined with classification algorithms (k-nearest neighbor, KNN; support vector machine, SVM; convolutional neural network, CNN) to recognize a rub-impact fault and its location. After comparative analysis with typical method, the result indicates that the proposed HE-AMW-MF method can identify and locate the rub-impact fault more precisely, and the precision is 32.0% and 36.2%, 22.5% higher than typical comparison method.

Making transformer hear better: Adaptive feature enhancement based multi-level supervised acoustic signal fault diagnosis

Acoustic signal fault diagnosis has been receiving increasing attention in the field of engine health management due to its effectiveness and non-invasiveness. Despite the progress made in fault diagnosis models, challenges still exist due to the complexity of acoustic signals and environmental factors. (1) End-to-end deep networks for fault diagnosis are at risk of underperformance or overfitting due to complex models and imbalanced data. (2) The complex acoustic environment within the vehicle power compartment poses obstacles to extracting subtle fault features. (3) Time–Frequency (TF) analysis has been proven to be an effective tool for characterizing the nonlinear features of fault signals, but it falls short in achieving ideal fidelity and resolution. To address these issues, an engine acoustic signal fault diagnosis method based on multi-level supervised learning and time–frequency transformation was proposed. First, adopting a multi-level supervised learning paradigm decomposes the fault diagnosis task into three stages: feature enhancement, fault detection, and fault identification, thereby incorporating additional experiential knowledge to mitigate overfitting. Second, an adaptive fault feature band extraction algorithm based on the fusion of multiple time–frequency analyses is proposed, specifically for extracting unique features from different vehicle datasets. Finally, a frequency band attention module was designed to focus on the frequency range most relevant to the characteristics of engine fault. The proposed method was validated on various audio signal fault datasets, and the results indicated its superior performance compared to other state-of-art fault detection and identification methods.

Audio-Based Engine Fault Diagnosis with Wavelet, Markov Blanket, ROCKET, and Optimized Machine Learning Classifiers.

Engine fault diagnosis is a critical task in automotive aftermarket management. Developing appropriate fault-labeled datasets can be challenging due to nonlinearity variations and divergence in feature distribution among different engine kinds or operating scenarios. To solve this task, this study experimentally measures audio emission signals from compression ignition engines in different vehicles, simulating injector failures, intake hose failures, and absence of failures. Based on these faults, a hybrid approach is applied to classify different conditions that help the planning and decision-making of the automobile industry. The proposed hybrid approach combines the wavelet packet transform (WPT), Markov blanket feature selection, random convolutional kernel transform (ROCKET), tree-structured Parzen estimator (TPE) for hyperparameters tuning, and ten machine learning (ML) classifiers, such as ridge regression, quadratic discriminant analysis (QDA), naive Bayes, k-nearest neighbors (k-NN), support vector machine (SVM), multilayer perceptron (MLP), random forest (RF), extra trees (ET), gradient boosting machine (GBM), and LightGBM. The audio data are broken down into sub-time series with various frequencies and resolutions using the WPT. These data are subsequently utilized as input for obtaining an informative feature subset using a Markov blanket-based selection method. This feature subset is then fed into the ROCKET method, which is paired with ML classifiers, and tuned using Optuna using the TPE approach. The generalization performance applying the proposed hybrid approach outperforms other standard ML classifiers.

Model-Assisted Probabilistic Neural Networks for Effective Turbofan Fault Diagnosis

A diagnostic method for gas-path faults of turbofan engines, relying on a Probabilistic Neural Network (PNN) coupled with a thermodynamic model of the engine, is presented. The novel aspect of the method is that its training information is generated dynamically by an accompanying Engine Performance Model. In the proposed approach, the PNN efficiently addresses the first step of a diagnostic process (i.e., detection of the faulty component at the current operating point), while with the aid of an adaptive engine model, the fault is then further isolated and identified. A description of the proposed method and training aspects of the PNN are presented. The method is applied to the case of a mixed-flow turbofan engine to diagnose common gas-path faults in compressors and turbines (i.e., fouling, FOD, erosion, and tip clearance). Its performance is evaluated using realistic fault data that may be acquired at various operating conditions within a flight envelope.

Intelligent Helicopter Turbine Engine Fault Diagnosis Using Multi-Head Attention

A turbine engine provides power to the helicopter, enabling the helicopter to travel and hover in the air. Since the rotorcraft operates at high altitudes, ensuring safety and maintaining a healthy operational status are crucial at all times. Therefore, a prognostics and health management (PHM) system for the turbine engine must be implemented to predict any anomalies or faults to prevent catastrophic accidents. This research proposes a novel fault diagnosis method for helicopter turbine engines based on operational data acquired from actual aircraft. First, the proposed method predicts engine torque using other operational data while accounting for uncertainty. A Bayesian regression approach is employed to predict the engine torque. The torque margin, defined as the difference between the actual torque and the estimated torque, is then used to diagnose engine faults. Specifically, a multi-head attention mechanism is incorporated to capture interactions between various engine parameters. Additionally, domain adaptation techniques are applied to enhance the model's generalization performance, ensuring robustness across diverse operating conditions. The proposed method is validated using seven different datasets, each acquired from a helicopter engine. Four datasets were used for training, while the remaining three were allocated for testing and validation. The results indicated that the proposed method accurately predicted torque. Furthermore, the fault diagnosis showed promising results, leading to a 3rd-place finish in the 2024 PHM Society Data Challenge in terms of validation score.

Data-driven Detection of Engine Faults in Infrequently-driven Ground Vehicles

We investigated the detection of engine faults in infrequently driven ground vehicles using data-driven methods based on neural network autoencoders. Multivariate time-series data from the infrequently driven vehicles under investigation had limited coverage of operating conditions. Hence, a considerable part of this work focused on identifying suitable vehicles, relevant signals, and pre-processing the data. We trained autoencoder models on eight vehicles with known faults and detected faults in six. Four of the faults were detectable under idle conditions and four were detectable under driving conditions. Model evaluations required human inspection to distinguish fault detections from other anomalies. We detail our procedures for pre-processing, model development, and post-processing, and we include a discussion on our interpretations of the model results.

Fault detection method of new energy vehicle engine based on wavelet transform and support vector machine

In order to solve the problem that existing methods are affected by fundamental frequency and harmonic frequency oscillation of new energy vehicle engine, a new energy vehicle engine fault detection method based on wavelet transform and support vector machine is proposed. Firstly, a detection model of abnormal noise signal of automobile engine fault is established, and the time-frequency parameters of basis function are adjusted adaptively. Then, the mechanical excitation component and the battery excitation component in the engine surface radiation noise are separated, and the new energy vehicle engine fault signal is decomposed by feature decomposition and multi-scale separation. Finally, wavelet transform combined with support vector machine algorithm was used to extract fault features of new energy vehicles, fuzzy clustering was carried out, and time-frequency analysis of fault signals was carried out in fractional Fourier domain to realize fault detection of new energy vehicle engines. The test results show that the method has high fault identification level and good clustering of fault feature points for new energy vehicle engine fault detection, and can effectively improve the ability of engine fault detection with good application effect.

Internal Combustion Engine Fault Detection Based on Random Convolutional Neural Networks

The internal combustion engine plays a very important role in many fields, and when the internal combustion engine fails, if it is not found in time, it may cause continuous damage to the internal combustion engine, and further affect the life of the entire mechanical system. To solve the above problems, this paper proposes fault detection of internal combustion engines based on a random convolutional neural network. The cylinder burst noise signal of the internal combustion engine is decomposed by a stationary wavelet, and the inclusion matrix is composed of partial decomposition coefficients. Using singular value theory, the singular value containing matrix is extracted as the characteristic of cylinder burst noise signal. The singular value is used as the input of a random convolutional neural network to train and identify faults. The AdaBound optimizer was then applied to adapt the learning rate to changes, thereby accelerating the weight update of the model. At the same time, the neurons in the structure are randomly deactivated by dropout technology to prevent complex collaborative responses to the training data, and the diagnosis results of each network model are integrated by Dempster synthesis rules. The experimental results show that the minimum mean square error of 0.00165 is achieved when there are 15 neurons in the hidden layer, and it gradually increases as the number of neurons increases. Therefore, it is determined that there should be 15 neurons in the hidden layer, and the trainLM algorithm is used. The decision factors for training, validation, cross-validation, and overall data are 0.98947, 0.98597, 0.97738, and 0.9801, respectively, all reaching the control accuracy of 0.95, which indicates that the random convolutional neural network proposed has a high accuracy for internal combustion engine fault detection. For the main bearing Y-directional force, as the gap increases, its fluctuation trend strengthens; the main bearing force changes within a small range at first, then increases sharply when the gap reaches 0.2[Formula: see text]mm, and some main bearings even experience significant impact loads.

Analysis and optimization of abnormal noise in lubricating oil circuit of diesel engine

Engine abnormal noise is one of the common engine faults, which will affect the comfort, power and safety of the automobile at the same time. In order to study the abnormal noise existing in the test process of diesel engine lubricating oil circuit system, 1D numerical simulation analysis of its flow field is carried out by using simulation software GT-power to verify the boundary conditions of the model and analyze the fluctuation of pressure at different speeds and temperatures. Through numerical analysis, it is found that there is almost no pressure fluctuation in the lubrication system before the pressure limiting valve is opened, but after the pressure limiting valve is opened, pressure fluctuation occurs in the pipeline, and the pressure fluctuation is different at different positions. Finally, the lubrication pipeline is simulated and analyzed by Pumplinx software, and the pressure fluctuation in the lubrication pipeline of diesel engine is reduced by optimizing the diameter of the oil pipeline and increasing the cavity structure.

Aircraft engine fault diagnosis based on fused convolutional Transformer

Aircraft engines will face corrosion and erosion problems when working in harsh gas path environments for a long time, and the fault parameter characteristics are not obvious. Therefore, accurate aircraft engine fault diagnosis methods are of great significance to ensure the safe operation of aircraft. In order to improve the prediction accuracy, an aircraft engine fault diagnosis method based on fused convolutional Transformer is proposed. The self-attention mechanism is used to extract useful features and suppress redundant information. The maximum pooling layer (MaxPool) is introduced into the Transformer model to further reduce the model memory consumption and parameter quantity, and alleviate the overfitting phenomenon. The simulation data set of turbofan engines based on GasTurb modeling is used for verification. The results show that compared with the Transformer network and other traditional deep learning models Back Propagation Neural Networks (BP network), Convolutional Neural Networks (CNN) and Recurrent Neural Network (RNN), the accuracy rates are improved by 6.552%、28.117%、13.189%and respectively 10.29%, which proves the effectiveness of the proposed method and can provide a certain reference for aircraft engine fault diagnosis.

A Review of Aircraft Engine Fault Prediction and Performance Optimization Based on Deep Learning

This paper is aimed at presenting a systematic review of the recent studies on the use of deep learning techniques in aircraft engine health management for fault detection and performance enhancement. This comes in the light of increased growth in the air transport industry, the efficiency of aircraft engines and the efficiency of their maintenance have emerged as key issues in the safety and efficiency of air transport. This paper also identifies some of the key issues emerging from this area such as the data issues and the need to develop very accurate models for faulting and performance enhancement. Some of the most advanced techniques, namely the Long Short-Term Memory (LSTM) networks and Convolutional Neural Networks (CNNs) are discussed as potential solutions to these challenges. The theoretical descriptions of the working principles, as well as the strengths and weaknesses of the methods in question with regard to the problems of aircraft engine health management, are elaborated. Furthermore, this paper introduces the application of Reinforcement Learning as a new technique in the area of optimal control of engine parameters with the hope of improving the dependability and maintainability of aircraft engines and in the process of achieving sustainable growth in the aviation industry. The issues on model interpretability, model generalization and model deployment are also discussed and research guidelines for the future work in this field are also outlined. In this paper, the state of the art and the importance of deep learning as well as the applications of deep learning in aircraft engine health management are discussed based on the literature and case studies.

Fault Detection on Short-Haul or Highly Dynamic Flights Using Transient Flight Segments

Abstract A machine learning-based approach is presented, which allows to detect persistent engine faults after a single flight. It utilizes transient in-flight measurements and a transient engine model. The time series of the residuals between the measured data and the data resulting from performance synthesis is evaluated using moving windows containing at least one transient segment. A continuous wavelet transformation and a pretrained convolutional neural network are utilized on the residuals for feature extraction. The fault detection is carried out via a one-class support vector machine, trained exclusively on nominal engine operation data. Therefore, the approach requires no a-priory knowledge of the effects of engine faults on the in-flight measurements. Under the assumption of persistent faults, all windows of a single flight, which contain at least one transient segment, are considered in order to improve the reliability of the fault detection. This approach is validated using measured data of a small helicopter engine that replicates the dynamic flight of the corresponding model helicopter on a ground test bed. Consequently, step changes as well as complex variations of the shaft power output are considered. Four standard gas path sensors are considered. The one-class support vector machine is used successfully to detect two types of total pressure sensor anomalies. Assuming a typical number of transient segments for an average short haul flight, it turns out that persistent faults can be detected within one flight with a probability of above 90%.

Improved Simplified Engineering Fault Displacement Hazard Evaluation Method for On-Fault Sites

The safety of high-potential risk facilities concerned with fault displacement hazards is a complex technical issue, especially if the fault is revealed beneath the facility during the operation. Applying simple conservative engineering hazard evaluation methods is rational if an urgent decision should be made to continue operation or implement protective measures. Engineering methods are being published for strike-slip on-fault sites and structures. Their crucial point is to estimate the probability of the rupture at the site intersection and consider the displacement distribution over the rupture length relative to the site’s on-fault location. It is shown in the paper that strict geometrical relations between the site location, length and initial point of the rupture determine whether the principal fault displacement intersects the site. The paper considers these geometrical parameters as independent random variables and applies a screening of ruptures contributing to the hazard. Magnitude- and on-fault coordinate-dependent empirical relations have been analysed and selected to evaluate the site displacements. The procedure resulted in realistic but conservative hazard curves for different on-fault site locations using data from the Paks site in Hungary. The results were compared to those obtained by the conservative engineering method for the same site and some published analyses.