Next Article in Journal
Life Cycle Assessment of Torrefied Residual Biomass Co-Firing in Coal-Fired Power Plants: Aspects of Carbon Dioxide Emission
Next Article in Special Issue
Effects of Two-Stage Injection on Combustion and Particulate Emissions of a Direct Injection Spark-Ignition Engine Fueled with Methanol–Gasoline Blends
Previous Article in Journal
Heat Transfer Mechanism Study of an Embedded Heat Pipe for New Energy Consumption System Enhancement
Previous Article in Special Issue
Discovery and Analysis of Key Core Technology Topics in Proton Exchange Membrane Fuel Cells Through the BERTopic Model
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

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

by
Norah Nadia Sánchez Torres
1,
Jorge Gomes Lima
2,
Joylan Nunes Maciel
1,
Mario Gazziro
3,
Abel Cavalcante Lima Filho
4,
Cicero Rocha Souto
2,
Fabiano Salvadori
2,* and
Oswaldo Hideo Ando Junior
1,2,5,6,*
1
Interdisciplinary Postgraduate Program in Energy & Sustainability (PPGIES), Federal University of Latin American Integration—UNILA, Av. Tancredo Neves, 3147, Foz do Iguaçu 85867-000, PR, Brazil
2
Smart Grid Laboratory (LabREI), Center for Alternative and Renewable Research (CEAR), Federal University of Paraiba (UFPB), Jardim Universitário, s/n, João Pessoa 58051-900, PB, Brazil
3
Information Engineering Group, Department of Engineering and Social Sciences (CECS), Federal University of ABC (UFABC), Av. dos Estados, 5001, Santo André 09210-580, SP, Brazil
4
Department of Mechanical Engineering (DEME), Technology Center (CT), Federal University of Paraiba (UFPB), Jardim Universitário, s/n, João Pessoa 58051-900, PB, Brazil
5
Research Group on Energy & Energy Sustainability (GPEnSE), Academic Unit of Cabo de Santo Agostinho (UACSA), Federal Rural University of Pernambuco (UFRPE), Rua Cento e Sessenta e Três, 300, Cabo de Santo Agostinho 54518-430, PE, Brazil
6
Program in Energy Systems Engineering (PPGESE), Academic Unit of Cabo de Santo Agostinho (UACSA), Federal Rural University of Pernambuco (UFRPE), Rua Cento e Sessenta e Três, 300, Cabo de Santo Agostinho 54518-430, PE, Brazil
*
Authors to whom correspondence should be addressed.
Energies 2024, 17(23), 6164; https://doi.org/10.3390/en17236164
Submission received: 19 October 2024 / Revised: 26 November 2024 / Accepted: 3 December 2024 / Published: 6 December 2024
(This article belongs to the Special Issue Optimization of Efficient Clean Combustion Technology)

Abstract

:
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.

1. Introduction

Global greenhouse gas (GHG) emissions, especially in the energy and transportation sectors, have been steadily increasing. Since 1990, global emissions have grown by 41%, with the energy sector responsible for 73% of these emissions. Among them, the transportation sector alone contributes 15% of total emissions, primarily due to fossil fuel consumption in internal combustion engines [1,2,3].
In 2019, a study by Det Norske Veritas (DNV) found that the 25,000 largest ships, 30% of the global fleet, were responsible for 80% of CO2 emissions. A total of 93% of the global fleet continues to run on conventional fossil fuels [4]. The IMO’s Fourth GHG Study of 2020 estimated that GHG emissions from shipping in 2018 accounted for around 2.89% of global anthropogenic GHG emissions and that such emissions could account for between 90% and 130% of 2008 emissions by 2050 [5]. IMO’s clear targets are large-scale decarbonization by or around 2050, a 20% reduction in emissions by 2030 and a 70% reduction by 2040 [6].
In this context, Brazil predominantly relies on a fleet powered by internal combustion engines, underscoring the importance of seeking alternatives and adopting technologies to promote sustainability in the transportation sector aligned with the Sustainable Development Goals (SDGs) of Agenda 2030 and the commitments of the Paris Agreement for achieving net-zero emissions by 2050. Moreover, Brazil has committed to reducing CO2 emissions by 37% by 2025, with a target of 43% reduction by 2030 [1,2,3].
The rapid development of digital technology has brought dramatic changes to many sectors, including the world of internal combustion engines essential for transportation, energy production, and manufacturing.
Internal combustion engines operate by converting chemical energy from the reaction between fuel, oxidizer, and heat into mechanical energy through combustion in a confined space, generating energy to move vehicles and drive machinery [7]. These engines, commonly used in transportation and industry, operate based on the principles of thermodynamics, primarily by altering the pressure and temperature of the system. The combustion of the fuel triggers a rapid expansion of gases to drive the pistons in the cylinders, which in turn rotate a crankshaft to produce mechanical movement [8,9,10].
However, the demand for greater economy, lower emissions, and higher performance requires the use of advanced engineering, design, and operational management techniques [11,12]. Digital twins alleviate these types of problems by providing a complete platform for holistic engine analysis, allowing researchers to gain insights into complex behaviors and interconnections that are often complex or expensive to investigate experimentally [13].
The incorporation of digital twin technology into the domain of internal combustion engines has created new opportunities for innovation and optimization. A digital twin is a virtual clone of a physical thing, in this case an internal combustion engine, that operates alongside its physical counterpart [14]. It involves a virtual replica that accurately represents the geometry, dynamics, and behavior of the engine in the digital world [15]. As a result of this symbiotic relationship, real-time data synchronization is possible, enabling continuous monitoring, evaluation, and simulation [13,16].
In light of these challenges, the development and implementation of advanced techniques using non-invasive sensors are crucial for the widespread adoption of prediction and fault identification systems in internal combustion engines.
Non-invasive techniques are crucial for analyzing failures in internal combustion engines, as they do not directly interfere with the system, operating externally, without direct contact or disassembly, which reduces the risk of damage to the equipment or engine.
The most frequently investigated failures include mechanical problems, lubrication system failures, and cooling system failures. Mechanical failures can include valve clearance problems, ignition failures, and fuel injection difficulties. Lubrication system failures are often associated with variations in oil pressure and viscosity, while cooling system failures include water leaks and engine overheating [10,17,18].
To mitigate the impacts of these failures, different pre-injection parameters are used, which are analyzed using techniques that promote the reduction of nitrogen oxide (NOx), carbon monoxide (CO), and hydrocarbon (HC) emissions [19,20,21,22].
In this context, the adoption of sensors and transducers becomes essential for the diagnosis and monitoring of failures in real time. Some authors highlight that vibration and acoustic measurements are among the most effective non-invasive or non-destructive techniques used to diagnose faults [19,23,24,25,26]. For vibration measurements, accelerometers are widely used, they are instruments that monitor changes in speed and acceleration due to shocks, vibration, or impact, being applicable in several areas, including the nautical, aeronautical, and aerospace industries. The main types of accelerometers include piezoelectric, capacitive, inductive, potentiometric, and piezoresistive. Microphones are also used as acoustic data acquisition devices, converting acoustic energy resulting from mechanical sound waves into electrical energy [26,27,28,29].
On the other hand, lubricating oil diagnostics can be performed using sensors installed in the lubrication system, allowing continuous monitoring of oil conditions without the need for manual collection or the interruption of engine operation. Exhaust gas analysis diagnostics can also be performed using sensors, using an exhaust gas analyzer with simultaneous recording of the load indicator, engine speed, inclinometer and Global Positioning System (GPS) data, or with a standard (stationary) measurement system and a portable diagnostic system [18,26].
In the diagnosis of faults in internal combustion engines, Artificial Intelligence (AI) is used because it offers the ability to process large volumes of complex data from sensors efficiently and in real time. Methods such as artificial neural networks are used due to their ability to grasp and recognize complex patterns even in noisy data. In addition, ANNs can adapt to different operating conditions, providing personalized diagnostics and fault predictions before they impact engine operation [30,31]. These technologies combine precision, cost reduction, efficiency, and flexibility, making them essential tools for the advanced monitoring of internal combustion engines.
Classification methods, such as Probabilistic Neural Networks (PNN), Back Propagation, and Support Vector Machines (SVMs), are used to improve diagnostic accuracy, allowing a more effective classification of the processed signals [32,33]. In addition, the Deep Neural Network (DNN) method coupled with Virtual Sample Generation (VSG) is used to establish engine performance models [34]. Signal processing techniques, such as Fast Fourier Transform (FFT) and Discrete Wavelet Transform (DWT), are employed to process the signals obtained from vibration and acoustic sensors, filtering noise and highlighting anomalies related to engine performance [29,35].
These technologies aim not only to maximize energy efficiency and reduce operational costs but also play a crucial role in reducing GHG emissions. Effective data management and efficiency improvements enable enhanced operational and maintenance efficiency, prolonging equipment lifespan and mitigating pollutant emissions through optimized real-time operations. This approach not only aligns with global climate mitigation commitments but also drives the transition towards a more sustainable and resilient future.
Against this backdrop, this article presents a systematic review aimed at providing a comprehensive mapping of the state of the art in non-invasive techniques and the quest for open-access datasets for prediction and fault diagnosis in internal combustion engines. Motivated by mapping the state of the art in non-invasive techniques in internal combustion engines using the Proknow-C and Methodi Ordinatio methods to efficiently systematize and categorize relevant scientific literature, the study enables rigorous and structured literature selection based on the relevance and credibility of databases, resulting in a structured bibliographic portfolio that provides a solid foundation for the analysis of these techniques [36,37,38,39].
Furthermore, the analyses contained in this systematic review contribute not only to identifying gaps and opportunities for future developments but also to driving advancements towards the integration of AI algorithms with digital twins to identify gaps for further improvements in fault diagnosis and prediction techniques. The paper’s contributions are summarized as follows:
Systematic Review Applying Proknow-C and Methodi Ordinatio Methods: An innovative approach to fault research in internal combustion engines demonstrating the effectiveness of combining these methods to efficiently systematize and categorize relevant scientific literature, resulting in a structured bibliographic portfolio that offers a solid foundation for significant insights and future research and development in the field.
Mapping of Non-Invasive Techniques: The study identifies and analyzes non-invasive methods for detecting faults in internal combustion engines, emphasizing their importance in minimizing interference in monitored systems and ensuring engine integrity.
Mapping and Integration of Artificial Intelligence (AI) with Digital Signal Processors (DSPs): An analysis of AI algorithms used for fault diagnosis, demonstrating the flexibility and effectiveness of combining various acquisition systems (such as sound, vibration, temperature, and current) with advanced AI and DSP techniques for comprehensive and efficient fault diagnostics.
Mapping and use of Digital Twins: An analysis of digital twins for fault detection, demonstrating that they provide an efficient, economical, and safe way to manage and maintain internal combustion engines, contributing to a more reliable and sustainable operation.
Finally, the article is structured into four sections addressing different aspects discussed. Section 1 provides a brief introduction to the importance of non-invasive techniques for prediction and fault diagnosis in internal combustion engines and their contribution to energy sustainability. Section 2 details a systematic review using the Proknow-C and Methodi Ordinatio methods, enabling rigorous and structured literature selection, mapping the state of the art in studies, experiments, and the availability of open datasets for the applicability of these techniques. Section 3 discusses the key findings and advancements resulting from the systematic review, emphasizing the significant contributions of non-invasive techniques to fault diagnosis and detection in internal combustion engines. Finally, Section 4 offers concluding remarks and suggestions for future research, exploring potential developments based on the integration of artificial intelligence and digital twins for internal combustion engines in different scenarios, identifying gaps and opportunities for future research.

2. Systematic Review

This section describes the process used to search, analyze, and select the set of relevant scientific research on the topic, which is used as the theoretical basis, called the bibliographic portfolio (BP). The literature review was conducted on the research topic “Naval Telemetry platform for coastal waterway transport” and “Digital Twin in fault detection in internal combustion engines”, applying the Proknow-C [38] and Ordinatio [40] methods to obtain a well-structured BP without gaps in the searches, allowing for a more rigorous selection of relevant literature, thereby increasing the quality and credibility of the results. Both methods allow for systematic bibliographic research and a structured and efficient identification of a BP with the most relevant scientific articles on the research topic.
The Proknow-C methodology encompasses several rigorous stages, including the selection of a portfolio of relevant articles, bibliometric analysis, and a systematic evaluation of the bibliographic portfolio. Bibliometric analysis assesses characteristics such as citation frequency, influential authors, and trends in the field. Systematic analysis considers theoretical aspects, methodologies, and the results found in the selected studies. By applying the Proknow-C method, it is possible to establish a solid knowledge base on the subject, contributing to supporting and guiding research in a more efficient and accurate way [36,39].
Methodi Ordinatio is a systematic review methodology that guides the search, selection, collection, and classification of scientific articles, using ICTs as support, with some improvements applied, such as the use of JabRef to collect data from scientific articles. What sets Methodi Ordinatio apart from other systematic review methodologies is the use of the InOrdinatio equation, which allows articles to be classified according to their scientific relevance. The equation works with the three most important factors in a scientific article: the impact factor, the year of publication of the research, and the number of citations [37].
Figure 1 and Figure 2 show the process of each approach, respectively.
Figure 3 shows the process carried out to classify the order of filtering the databases from the raw articles to obtain the BP: (i) the removal of duplicates; (ii) the removal of articles with titles misaligned with the project; (iii) the identification of impact factor, originating repository A; (iv) scientific recognition, originating repository B; (v) the removal of articles with abstracts misaligned with the project, originating repository C; and (vi) the full reading and removal of articles misaligned with the theme, originating the BP.
Through the proposed method, the keywords (Kws) were defined from the research axes described below for the first proposed theme (“Naval telemetry platform for coastal waterway transport”). The three research axes, (a) non-invasive methods, (b) fault detection, and (c) internal combustion engines, were defined knowing that the state of the art must include techniques for fault detection to resolve the following points: the application of a new technique to identify faults in combustion engines with high sensitivity to initial conditions and low computational effort; the development of an algorithm for identifying ignition and combustion faults in combustion engines; and the diagnosis of ignition, combustion, and engine faults, considering non-invasive methods.
After defining the Kws for each axis, combinations were elaborated based on the logical expressions “Axis 1 and Axis 3”, “Axis 2 and Axis 3”, and “Axis 1 and Axis 2 and Axis 3”, searched on 25 May 2024. Only articles in English published in qualified journals and up to 20 years old were considered. Theses, dissertations, and conference papers were disregarded to ensure a high level of quality, validity, and reliability.
Next, a search was conducted in three academic article databases using the Publish or Perish® software Version: 8.16.4790 (20 October 2024): SCOPUS, Web of Science (WoS), and Google Scholar. The search results are described in Table 1 and Figure 4.
The search results were exported to Zotero® software Version: 7.0.10 (18 November 2024), a reference manager, containing a raw article base of 1480 articles. Applying the first and second filters of the methodology used, 287 articles with titles aligned with the project theme and without duplicates were obtained. Analyzing the remaining articles for impact factor, filtering by date and most cited, 51 articles were added to repository A. For the remaining articles, the analysis was based on scientific recognition, resulting in 29 articles, added to repository B. After reading the abstracts of the articles in repositories A and B, 30 articles proved relevant and were added to repository C.
To conduct the full reading of the articles in repository C, articles not freely available were excluded, along with articles analyzed and found irrelevant to the theme. Thus, the BP is composed of 19 articles. Table 2 presents the BP, ordered by the number of citations, and Figure 5 graphically shows the entire process.
In this subsection, using the proposed method, the definition of the keywords (Kws) was carried out based on the research axes described below for the second proposed theme (“Digital Twin in fault detection in internal combustion engines”), knowing that the state of the art must include fault detection techniques to resolve the points mentioned in the previous theme. Axis 4 is added, a Digital Twin, which is mixed with the three axes of the first theme (non-invasive methods; fault detection; and internal combustion engines).
After defining the Kw, the combinations were created based on the logical expressions Axis 4 and Axis 3, Axis 4 and Axis 3 and Axis 2. It is worth noting that research was carried out with the inclusion of Axis 1, and that no results were found, searching on the same date and considering the same parameters as the previous theme.
Then, a search was carried out in three academic article databases, using the Publish or Perish® software. These databases included SCOPUS, Web of Science (WoS), and Google Scholar. The results of the research carried out are described in Table 3.
The search results were exported to the Zotero® software, a bibliographic reference manager, containing a raw article base of 1310 articles. Applying the six filters, 12 articles were obtained that are part of the digital twin BP. Table 4 shows the BP, ordered by the number of citations.

3. Results and Discussions

Analyzing the articles in the bibliographic portfolio (BP), it is observed that the majority of important articles (eleven articles) are from the last 5 years, with 2019 having the highest number of relevant publications (four articles) for this research. Additionally, Poland is the country with the most publications, as shown in Figure 6 and Figure 7, respectively.
In [33,41,54], the means of signal acquisition are sound and vibration, used to identify mechanical damage in internal combustion engines (ICEs), and the methods adopted for signal processing differ.
In the first study [41], vibration measurements were carried out by PCB piezotronics accelerometers placed at four points on the engine head surface, and the received signals were processed using Discrete Fourier Transform (DFT). Acoustic signals were captured using a spectrum analyzer, random incidence microphone, condenser preamplifier, and a calibrator, processed by Fast Fourier Transform (FFT). The collected signals can be used to detect phenomena in the combustion chamber, indicating ignition failures.
The study [33] utilizes Discrete Wavelet Transform (DWT) and Probabilistic Neural Networks (PNN); tests were performed using a multichannel recording device where vibratory signals were recorded at a frequency of 50 kHz, measured by vibration acceleration transducers. The recorded signals were pre-processed using DWT, allowing the extraction of the correct signal from random noise and enabling the development of a real-time operating system. Various DWT types were tested, and the Meyer Wavelet proved most advantageous. The PNN consists of three layers (input, hidden, and output), with experiments dividing the data into 50% training and 50% testing groups. Depending on the data pattern, the network can have two to eleven inputs, with 100 neurons in the hidden layer, resulting in four outputs. Consequently, neural classifiers were developed that function perfectly or nearly perfectly, distinguishing mechanical damage in engines.
The third study [54] presents a more current procedure focused on marine ICEs, proposing a system for reliability identification from sound and vibration signals collected in each cylinder of the engine using machine learning. The innovation in this method is the increased case count and the implementation of the Support Vector Machine (SVM) classifier. The technique’s efficiency was tested for complete and incomplete learning datasets, resulting in 100% reliability identification in both cases.
Vibration is used as a signal acquisition method in [35,42,43,49,51,56], diagnosing valve clearance size, valve failure, various failure types, ignition failure, and both ignition and mechanical failures in ICEs, respectively. The methodology proposed by [42] is based on signal analysis using Artificial Neural Networks (ANNs). Vibrations were measured using two piezoelectric accelerometers mounted on the engine head casing. For the ANN, the classifier database was created by training the network from engine vibration signals with predefined valve clearances. With the classifier diagnosing correctly, the technique can be applied to both the training engine and other engines of the same type. A data acquisition system is presented in [35], comprising a single-axis ceramic shear accelerometer, sound and vibration input module, and an Ethernet module. Data are processed in MATLAB using FFT and the SVM algorithm, showing a diagnostic accuracy of 97%. The method in [49] proposes an automated system based on ANNs, specifically MLP and PNN, to diagnose a series of various types of ICE failures. The system automation was divided into three parts: detection, localization, and severity identification. The network was entirely trained from simulated models and tested in real experimental cases. The signal processing approach for diagnosis was based on envelope analysis of vibration signals captured by accelerometers on the engine block surface, demonstrating that networks trained through simulation can efficiently detect combustion and mechanical failures, and identify failure location and severity. The procedure in [51] bases vibration signal capture on a smartphone placed on the engine, processed by two different methods to test viability and computational effort: Wavelet Multiresolution Analysis (WMA) and chaos-based signal analysis using maximum density (SAC-DM) for ignition failure identification. Both techniques achieved a 100% failure diagnosis rate, with SAC-DM requiring less computational effort. The technique developed in [56] distinguishes itself by learning only from healthy sample data based on ICE vibration signals processed using FFT and DWT, classified using a One Class Support Vector Machine (OCSVM). Classification algorithms were tested on twelve real engines, showing that all ignition and mechanical failures could be detected with minimal false positives and an error rate of 0.15%. In [43], vibration capture is also used, presenting a state-of-the-art study where the FFT and Morlet Wavelet methods stand out.
The author of [67] addresses the development of a multi-scale multisensory signal fusion method for fault detection in high-speed, high-power diesel engines under variable operating conditions. The approach combines data collected by different sensors (such as vibration and pressure) and uses advanced machine learning techniques, including Convolutional Neural Networks, to improve diagnostic accuracy and robustness. The method is evaluated in experimental scenarios that simulate different bearing and engine faults, demonstrating its effectiveness under dynamic conditions.
In studies [25,46,50,55], acoustic signals are used to diagnose ICE conditions, differing in the processing methods developed. The first primarily uses a decision tree, tested with a sound pressure level meter. Thirteen acoustic wave spectrum measurements were performed for different induced engine failure cases, with the decision tree used for damage diagnosis from spectral sound emission characteristics. The second detects fuel injection system failures specifically in a marine ICE, with sound recorded in two locations, the injector pump outlet and the injector inlet, using a broadband accelerometer and a sound card. The amplitudes of vibratory pulses in a normally operating ICE change when a failure occurs, enabling the proposed system to detect malfunction. The third study uses ANN, DWT, and fractal dimension, integrating low-cost signal acquisition hardware (Arduino Due and electret condenser microphone CMA-4544PF-W) and processing software running on smartphones (Android). Fractal dimension application expedited analysis with low computational demand, distinct from DWT processing, with ANN applied for classification, achieving a 99.58% performance rate. The last proposes an acoustic characterization system using sound captured by a smartphone, processed by a deep learning method with a cascaded architecture, defined as conditional networks extracting sound for fault characterization. A multitask Convolutional Neural Network (CNN) predicts and characterizes the engine to detect ignition failures, achieving 87% diagnostic accuracy.
In [44,47,48], exhaust gas temperature is measured to detect ignition failures in the first and last studies, and combustion instability in the second.
The method in [44] uses a lambda sensor (O2) to capture temperature due to its low cost and computational effort, employing False Alarm Rate (FAR) and Success Detection Rate (SDR) techniques to avoid detection errors, as exhaust temperature fluctuates even under normal conditions. The sensor reading drops during ignition failure and rises upon normal combustion resumption, establishing diagnostic rules based on failure duration and sampling interval, resulting in an SDR of 0.75 and a fixed FAR, achieving 75% effectiveness.
An ultrasonic thermometry system is developed in [47], based on the thermal dependency of sound speed, with temperature measured by detecting the ultrasonic wave (USW) time-of-flight between the transmitter and receiver. This system offers high-speed response and an uncertainty rate below 0.73%, outperforming thermocouples.
In [48], temperature is measured using a laser-induced grating spectroscopy (LIGS) system, based on oscillation frequency measurement of the detected signal, less susceptible to noise effects. LIGS’s drawback is weaker signals at higher temperatures due to lower density and faster decay from rapid diffusion.
Seeking to reduce emission rates from an ICE, [45] develops a method to capture the oil film formed in the cylinder using ultrasonic imaging. A straight-beam ultrasonic contact transducer was used for wave propagation through the cylinder wall, collecting reflections as piston rings passed the detection area. The reflection coefficient varies with layer stiffness and corresponding acoustic properties of materials and lubricants, successfully diagnosing oil film thickness (OFT).
Vibration and pressure signals in the combustion chamber were analyzed in [52] to diagnose damage development in some marine ICE components, with data collection during natural vessel operation conditions during port maneuvers. Signals were processed by FFT, DWT, time–frequency domain analysis, and machine learning, used for ANN classification. The technique achieved an 80% result for fault identification and classification.
In [53], the fuel injector plays a crucial role in an ICE and is particularly susceptible to damage, directly affecting engine efficiency. Electrical properties of the injector coil can be collected from voltage and current curve analysis, determining injector health. A waveform analysis of the current by Derivative Sign Change (DSC) provides information on injector operation quality. Parameter changes are reflected in the initial waveform and DSC value and phase, verifying system status and detecting faults. Given the above, it is possible to visualize the methodology scheme used in each study for ICE fault diagnosis: first, the acquisition system is characterized, followed by the processing system and classification system for articles using Artificial Intelligence (AI).
Table 5 summarizes the research found in the BP, which initially showed that non-invasive methods are the most used for fault detection, although there are still 15% (three articles) that use invasive methods. In the BP articles, 73% (fourteen articles) use periodic sampling, 5% (one article) use real-time measurement, 11% (two articles) combine both forms, and 11% (two articles) use continuous measurement. Among the articles found, 47% diagnose mechanical faults, 21% diagnose ignition faults, 16% diagnose both faults and ignition, another 16% diagnose injection faults, and 5% diagnose oil film thickness.
This underscores the importance of using non-invasive methods for detecting faults in internal combustion engines. The table further shows that sound and vibration are the most commonly used acquisition systems for fault detection. Some articles even employ both acquisition systems together for fault detection, and there are also articles that utilize temperature, current, and cylinder oil systems.
It is important to note that articles using temperature, current, and cylinder oil acquisition systems employ Digital Signal Processor (DSP) processing systems, while others use AI-based processing systems, or a combination of both. This approach leverages the real-time efficiency and precision of DSPs alongside the adaptive capability and advanced analysis of AI, resulting in a more robust and efficient system for identifying faults in internal combustion engines.
Table 5 highlights that no article addresses CO2 diagnostics or monitors acquisition systems such as temperature, current, sound, vibration, fuel consumption, and overall quality control, nor does it monitor vessel data like location, displacement, route, and speed. Therefore, these would be unique aspects of the proposed development.
Table 6 presents a survey of BP articles utilizing AI, describing classification systems and indicating the datasets used for training. It is notable that out of nine sampled articles, seven employ various neural network architectures, offering a robust and flexible approach to diagnostic and predictive maintenance. Another parameter discussed is the sampling rate, ranging widely from 200 samples to 16,384 samples, highlighting the need for substantial and well-balanced datasets to accurately train, test, and validate AI models across a broad range of operational conditions and fault types.
Regarding the datasets presented in Table 6, many are proprietary, making them unavailable to the public. This difficulty in replicating results and independently validating models underscores another unique aspect of the proposed project: the development of a robust and comprehensive database that ensures appropriate partitioning for training, testing, and validation. This approach aims to achieve precise and replicable results, enabling the scientific community to validate and compare findings across different approaches.
In conclusion, Table 5 and Table 6 demonstrate a growing trend towards the use of non-invasive methods for fault detection, crucial for minimizing engine downtime and reducing maintenance costs. The combination of DSP and AI, validated by the articles in the tables, along with real-time and continuous monitoring, is essential for engine diagnostics and maintenance, addressing a gap identified in the literature. Another gap highlighted by both tables is the study of diverse parameters such as vibration, temperature, sound, current, and fuel consumption, among others, enabling a comprehensive and accurate analysis of internal combustion engine conditions and thus providing a robust and effective approach to fault diagnosis. Addressing the lack of publicly available data is also crucial for facilitating result validation and replication, as shown in the tables.
For the digital twin, analyzing the articles contained in the BP, it is possible to observe that all articles imported into the BP (12 articles) are from the last 5 years. In 2022, there was a greater number of published articles (five articles) relevant to the research, shown in the graph contained in Figure 8.
The use of real-time monitoring and testing is presented in [12,57,58], which demonstrate predictive insight through real-time data integration. In this case, digital twins have transformed the way an internal combustion engine is monitored during operation, providing a synchronous representation that simulates engine operation in real time, seamlessly integrating data from multiple sensors. They provide instant access to crucial information, including temperature, pressure, combustion efficiency, and emissions. In addition, any discrepancies between real and simulated processes are identified, suggesting potential problems or performance anomalies.
For anomaly detection and predictive analysis, real-time digital twins are essential to discover anomalies that may go unnoticed until maintenance or outages occur. As [14,65] show, digital twins can detect even the smallest differences by constantly comparing real data with predictions from virtual models. Furthermore, they integrate historical data and machine learning algorithms to provide predictability to digital twins, allowing them to predict likely errors based on patterns and trends.
For health monitoring and assessment, [63] presents how digital twins provide the user with a complete view of the engine status, as well as relevant information about the engine condition, thus allowing the evaluation of engine performance, providing real-time data and simulated performance. In [64], it is highlighted that this allows informed decisions about maintenance routines, component replacement, and performance optimization strategies to be made. Furthermore, it concludes that remote engine monitoring allows for more effective troubleshooting and decision-making, especially for engines located in remote or hazardous areas.
To improve data-driven decision-making, [62] presents that instead of relying on reactive measurements, digital twin data enables informed decisions that can improve efficiency, extend component life, and reduce downtime. In [60,61], it is shown that this transition to proactive decision-making changes maintenance techniques from scheduled operations to condition-based operations, optimizing resource utilization and reducing operating costs.
The use of AI and machine learning brings interesting opportunities for internal combustion engines. By integrating AI and machine learning, as shown in [59,66], digital twins can improve predictability, since AI algorithms can increase the accuracy of predictions and provide meaningful insights into optimal engine performance by learning from historical data. The data and comments presented are summarized in Figure 9, with each of the themes.
From the articles presented, we must highlight that [14,58,61,62,66] are applied in the aviation or aircraft sector. There are others such as [60], which is applied to energy, and for the maritime sector we have [12,64,65], which study and analyze maintenance in real time. Others analyze data on temperature, pressure, combustion efficiency, and emissions, so this is a clear gap found in the research, as shown graphically in Figure 10. Furthermore, it was not possible to find an article or review on the use of AI with a digital twin applied to internal combustion engines in maritime transport, for fault detection.

4. Conclusions

This study conducted a systematic review of non-invasive techniques for fault identification in internal combustion engines, using the Proknow-C and Methodi Ordinatio methods to categorize and systematize scientific literature. We provided a pioneering review of non-invasive techniques for fault detection, resulting in a comprehensive bibliographic portfolio.
In summary, the systematic review enabled a comprehensive mapping of non-invasive techniques for fault identification in internal combustion engines, highlighting the following: (i) the use of a wide range of AI algorithms, demonstrating the flexibility and effectiveness of these methods for fault diagnostics; (ii) the prevalence of non-invasive methods, emphasizing the importance of techniques that minimize interference in monitored systems; (iii) the combination of diverse acquisition systems with advanced data processing techniques, essential for comprehensive and accurate diagnostics; (iv) the need for continuous real-time monitoring, indicating the demand for systems capable of providing data instantly for monitoring and fault detection; (v) the use of digital twins as a non-invasive technique for continuous monitoring, diagnosis and predictive analysis in real time, which increases the reliability of the systems; and (vi) the lack of public datasets, which limits the replication of the results and the cross-validation between different studies.
A detailed analysis of the results revealed significant advances in the application of artificial intelligence and digital signal processing for precise and efficient fault diagnosis. Thus, research on non-invasive techniques for fault diagnosis in internal combustion engines is crucial for the development of more efficient, reliable, and sustainable engines. Artificial intelligence and digital twins emerge as a powerful tool to enhance fault diagnosis and prediction, contributing to optimizing the operation of internal combustion engines, enabling real-time predictive management, and potential for energy efficiency optimization and emissions reduction. The gaps identified in this systematic review highlight the following recommendations for future research:
Multimodal data fusion: Explore advanced methods to combine and analyze data from different sources (e.g., multiple sensor data, maintenance records, and operational data) to enhance the accuracy of fault detection. This includes multimodal data fusion techniques, integrating structured and unstructured data to provide a comprehensive view of the engine’s condition.
AI for fault prognosis and maintenance planning: Develop predictive models using advanced artificial intelligence algorithms such as deep neural networks, support vector machines, and reinforcement learning algorithms to forecast faults. This includes developing systems that not only identify faults but also recommend specific preventive actions to optimize engine life and efficiency.
Multiple diagnostics systems: Develop integrated systems capable of monitoring and diagnosing multiple aspects of engine performance, including mechanical failures, ignition problems, and CO2 emissions. These systems should operate in real-time, providing immediate information to operators and management systems, enabling rapid and efficient corrective actions.
Public datasets: Encourage the creation and sharing of high-quality datasets that include standardized and well-documented test data. This facilitates democratizing research and enables the validation and comparison of different fault diagnostic approaches, promoting consistent and replicable advances in the field. Some relevant initiatives, such as global collaboration, encourage the creation and sharing of data in various sectors, including energy and transportation. This allows researchers to access standardized data, promoting the validation of techniques and comparison between studies. PhysioNet inspires the use of datasets for machine learning, which can be applied to combustion engines; the National Renewable Energy Laboratory—NREL, provides a database related to renewable energy and energy efficiency, and the same model could be adapted to collect and make available data on failures; and finally, the Marine Engine Database, under development, gathers data for marine engine diagnostics and best practices.
Finally, it underscores the ongoing importance of technological innovations in fault diagnosis in internal combustion engines to promote energy digitalization and energy sustainability.

Author Contributions

Conceptualization: N.N.S.T., J.G.L., J.N.M., A.C.L.F., F.S. and O.H.A.J.; methodology: N.N.S.T., J.G.L., J.N.M., C.R.S., M.G., A.C.L.F., F.S. and O.H.A.J.; validation: J.N.M., F.S. and O.H.A.J.; investigation and simulation: N.N.S.T. and J.G.L.; writing—original draft preparation: N.N.S.T., J.G.L., J.N.M., C.R.S., A.C.L.F., F.S. and O.H.A.J.; writing—review and editing: M.G. and O.H.A.J.; project administration: F.S. and O.H.A.J.; funding acquisition: F.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research was partially supported by the FACEPE agency (Fundação de Amparo a Pesquisa de Pernambuco) throughout the project with references APQ-0616-9.25/21 and APQ-0642-9.25/22. O.H.A.J. and F.S. were funded by the Brazilian National Council for Scientific and Technological Development (CNPq), grant numbers 407531/2018-1, 303293/2020-9, 405385/2022-6, 405350/2022-8 and 40666/2022-3, as well as the Program in Energy Systems Engineering (PPGESE) Academic Unit of Cabo de Santo Agostinho (UACSA), Federal Rural University of Pernambuco (UFRPE). N.N.T.S. was funded by the Federal University of Latin American Integration (UNILA).

Data Availability Statement

Not applicable.

Acknowledgments

The authors thank Jecel Mattos de Assumpção Jr.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. IEA; IRENA; UNSD; World Bank; WHO. Tracking SDG 7: The Energy Progress Report; WHO: Washington, DC, USA, 2024; p. 179.
  2. Sachs, J.; Kroll, C.; Lafortune, G.; Fuller, G.; Woelm, F. Sustainable Development Report 2022, 1st ed.; Cambridge University Press: Cambridge, UK, 2022; ISBN 978-1-00-921005-8. [Google Scholar]
  3. CODS. Índice ODS 2022 para América Latina y el Caribe; Centro de los Objetivos de Desarrollo Sostenible para América Latina y el Caribe: Bogotá, Colombia, 2023; Volume 10, p. 100. [Google Scholar]
  4. Ovrum, E.; Longva, T.; Leisner, M.; Bachmann, E.M.; Gundersen, O.S.; Helgesen, H.; Endresen, O. Energy Transition Outlook 2024—Maritime Forecast to 2050; DNV: Oslo, Norway, 2023; p. 73. [Google Scholar]
  5. IMO. 2023 IMO Strategy on Reduction of Ghg Emissions from Ships; MEPC 80/WP.12; International Maritime Organization: Oslo, Norway, 2023; p. 18. [Google Scholar]
  6. UNCTAD. Review of Maritime Transport 2024: Navigating Maritime Chokepoints, 1st ed.; Review of Maritime Transport Series; United Nations Research Institute for Social Development: Bloomfield, NJ, USA, 2024; ISBN 978-92-1-106592-3. [Google Scholar]
  7. Heywood, J.B. Internal Combustion Engine Fundamentals. In McGraw-Hill Series in Mechanical Engineering; McGraw-Hill: New York, NY, USA, 1988; ISBN 978-0-07-028637-5. [Google Scholar]
  8. Bosch, R. Manual de Tecnologia Automotiva; Tradução da 25a ed. Alemã; Edgard Blücher: São Paulo, Brazil, 2005; ISBN 978-85-212-0378-0. [Google Scholar]
  9. Brunetti, F. Motores de Combustão Interna; Editora Edgard Blucher: São Paulo, Brazil, 2022; Volume 1, ISBN 978-85-212-1293-5. [Google Scholar]
  10. Nahim, H.M.; Younes, R.; Nohra, C.; Ouladsine, M. Complete modeling for systems of a marine diesel engine. J. Marine. Sci. Appl. 2015, 14, 93–104. [Google Scholar] [CrossRef]
  11. Lim, K.Y.H.; Zheng, P.; Chen, C.-H. A state-of-the-art survey of Digital Twin: Techniques, engineering product lifecycle management and business innovation perspectives. J. Intell. Manuf. 2020, 31, 1313–1337. [Google Scholar] [CrossRef]
  12. Stoumpos, S.; Theotokatos, G.; Mavrelos, C.; Boulougouris, E. Towards Marine Dual Fuel Engines Digital Twins—Integrated Modelling of Thermodynamic Processes and Control System Functions. J. Mar. Sci. Eng. 2020, 8, 200. [Google Scholar] [CrossRef]
  13. Tran, V.D.; Sharma, P.; Nguyen, L.H. Digital twins for internal combustion engines: A brief review. J. Emerg. Sci. Eng. 2023, 1, 29–35. [Google Scholar] [CrossRef]
  14. Xu, Z.; Ji, F.; Ding, S.; Zhao, Y.; Zhou, Y.; Zhang, Q.; Du, F. Digital twin-driven optimization of gas exchange system of 2-stroke heavy fuel aircraft engine. J. Manuf. Syst. 2021, 58, 132–145. [Google Scholar] [CrossRef]
  15. Cheng, D.-J.; Zhang, J.; Hu, Z.-T.; Xu, S.-H.; Fang, X.-F. A Digital Twin-Driven Approach for On-line Controlling Quality of Marine Diesel Engine Critical Parts. Int. J. Precis. Eng. Manuf. 2020, 21, 1821–1841. [Google Scholar] [CrossRef]
  16. Minchev, D.; Varbanets, R.; Shumylo, O.; Zalozh, V.; Aleksandrovska, N.; Bratchenko, P.; Truong, T.H. Digital Twin Test-Bench Performance for Marine Diesel Engine Applications. Pol. Marit. Res. 2023, 30, 81–91. [Google Scholar] [CrossRef]
  17. Neumann, S.; Varbanets, R.; Minchev, D.; Malchevsky, V.; Zalozh, V. Vibrodiagnostics of marine diesel engines in IMES GmbH systems. Ships Offshore Struct. 2023, 18, 1535–1546. [Google Scholar] [CrossRef]
  18. Dong, F.; Yang, J.; Cai, Y.; Xie, L. Transfer Learning-Based Fault Diagnosis Method for Marine Turbochargers. Actuators 2023, 12, 146. [Google Scholar] [CrossRef]
  19. Rodríguez, C.G.; Lamas, M.I.; Rodríguez, J.D.D.; Caccia, C.; Politecnico di Milano, Italy. Analysis of The Pre-Injection Configuration in a Marine Engine through Several MCDM Techniques. Brod. Int. J. Nav. Archit. Ocean. Eng. Res. Dev. 2021, 72, 1–17. [Google Scholar] [CrossRef]
  20. Hountalas, T.D.; Founti, M.; Zannis, T.C. Experimental Investigation to Assess the Performance Characteristics of a Marine Two-Stroke Dual Fuel Engine under Diesel and Natural Gas Mode. Energies 2023, 16, 3551. [Google Scholar] [CrossRef]
  21. Korczewski, Z. Test Method for Determining the Chemical Emissions of a Marine Diesel Engine Exhaust in Operation. Pol. Marit. Res. 2021, 28, 76–87. [Google Scholar] [CrossRef]
  22. Bogdanowicz, A.; Kniaziewicz, T. Marine Diesel Engine Exhaust Emissions Measured in Ship’s Dynamic Operating Conditions. Sensors 2020, 20, 6589. [Google Scholar] [CrossRef] [PubMed]
  23. Varbanets, R.; Shumylo, O.; Marchenko, A.; Minchev, D.; Kyrnats, V.; Zalozh, V.; Aleksandrovska, N.; Brusnyk, R.; Volovyk, K. Concept of Vibroacoustic Diagnostics of the Fuel Injection and Electronic Cylinder Lubrication Systems of Marine Diesel Engines. Pol. Marit. Res. 2022, 29, 88–96. [Google Scholar] [CrossRef]
  24. Tharanga, K.L.P.; Liu, S.; Zhang, S.; Wang, Y. Diesel Engine Fault Diagnosis with Vibration Signal. J. Appl. Math. Phys. 2020, 8, 2031–2042. [Google Scholar] [CrossRef]
  25. Varbanets, R.; Fomin, O.; Píštěk, V.; Klymenko, V.; Minchev, D.; Khrulev, A.; Zalozh, V.; Kučera, P. Acoustic Method for Estimation of Marine Low-Speed Engine Turbocharger Parameters. J. Mar. Sci. Eng. 2021, 9, 321. [Google Scholar] [CrossRef]
  26. Deptuła, A.; Kunderman, D.; Osiński, P.; Radziwanowska, U.; Włostowski, R. Acoustic Diagnostics Applications in the Study of Technical Condition of Combustion Engine. Arch. Acoust. 2016, 41, 345–350. [Google Scholar] [CrossRef]
  27. Junior, E.M.D.S. Técnica de Diagnóstico de Falhas em Motores à Combustão Interna Utilizando Aprendizagem de Máquina. Ph.D. Thesis, Programa de PósGraduação em Engenharia Mecânica, Universidade Federal do Rio Grande do Norte, Natal, RN, Brazil, 2018. [Google Scholar]
  28. Spada, A.L. Microfones: Parte 1, Attack do Brasil. 2016. Available online: http://www.attack.com.br/artigos_tecnicos/mic_1.pdf (accessed on 1 September 2024).
  29. Cabral, M.A.L. Classificação Automatizada de Falhas Tribológicas de Sistemas Alternativos com o uso de redes Neurais Artificiais Não Supervisionadas. Ph.D. Thesis, Universidade Federal do Rio Grande do Norte Centro de Tecnologia Programa de Pós-Graduação Em Engenharia Mecânica—PPGEM, Natal, RN, Brazil, 2017. [Google Scholar]
  30. Abubakar, S.; Said, M.F.M.; Abas, M.A.; Samaila, U.; Ibrahim, A.A.; Ismail, N.A.; Narayan, S.; Kaisan, M.U. Application of Artificial Intelligence in Internal Combustion Engines—Bibliometric Analysis on Progress and Future Research Priorities. J. Balk. Tribol. Assoc. 2024, 30, 632–654. [Google Scholar]
  31. Ahmed, R.; El Sayed, M.; Gadsden, S.A.; Tjong, J.; Habibi, S. Automotive Internal-Combustion-Engine Fault Detection and Classification Using Artificial Neural Network Techniques. IEEE Trans. Veh. Technol. 2015, 64, 21–33. [Google Scholar] [CrossRef]
  32. Yang, M.; Chen, H.; Guan, C. Research on diesel engine fault diagnosis method based on machine learning. In Proceedings of the 2022 4th International Conference on Frontiers Technology of Information and Computer (ICFTIC), Qingdao, China, 2–4 December 2022; IEEE: New York, NY, USA, 2022; pp. 1078–1082. [Google Scholar]
  33. Czech, P.; Wojnar, G.; Burdzik, R.; Konieczny, Ł.; Warczek, J. Application of the discrete wavelet transform and probabilistic neural networks in IC engine fault diagnostics. J. Vibroeng. 2014, 16, 1619–1639. [Google Scholar]
  34. Zheng, H.; Zhou, H.; Kang, C.; Liu, Z.; Dou, Z.; Liu, J.; Li, B.; Chen, Y. Modeling and prediction for diesel performance based on deep neural network combined with virtual sample. Sci. Rep. 2021, 11, 16709. [Google Scholar] [CrossRef] [PubMed]
  35. Venkata, S.K.; Rao, S. Fault Detection of a Flow Control Valve Using Vibration Analysis and Support Vector Machine. Electronics 2019, 8, 1062. [Google Scholar] [CrossRef]
  36. Maciel, J.N.; Ledesma, J.J.G. Forecasting Solar Power Output Generation: A Systematic Review with the Proknow-C; IEEE Latin America: Medellin, Colombia, 2021; Available online: https://ieeexplore.ieee.org/abstract/document/9448544/?casa_token=DYVEDw1-58gAAAAA:-dJjLLPdqeIt1UtTUfreE_cr95UvccHg9blr6Ab2Ca1b6Vud1zz7Y87lWGsGCr99eItiZzsPBIs (accessed on 15 September 2024).
  37. Regatieri, H.R.; Ando Junior, O.H.; Salgado, J.R.C. Systematic Review of Lithium-Ion Battery Recycling Literature Using ProKnow-C and Methodi Ordinatio. Energies 2022, 15, 1485. [Google Scholar] [CrossRef]
  38. Lacerda, R.T.D.O.; Ensslin, L.; Ensslin, S.R. Uma análise bibliométrica da literatura sobre estratégia e avaliação de desempenho. Gest. Prod. 2012, 19, 59–78. [Google Scholar] [CrossRef]
  39. Leandro, P.G.M.; Salvadori, F.; Izquierdo, J.E.E.; Cavallari, M.R.; Ando Junior, O.H. The Advancements and Challenges in Organic Photovoltaic Cells: A Focused and Spotlight Review using the Proknow-C. Energies 2024, 17, 4203. [Google Scholar] [CrossRef]
  40. Pagani, R.N.; Kovaleski, J.L.; Resende, L.M.M.D. Avanços na composição da Methodi Ordinatio para revisão sistemática de literatura. Cionline 2018, 46, 1886. [Google Scholar] [CrossRef]
  41. Barelli, L.; Bidini, G.; Buratti, C.; Mariani, R. Diagnosis of internal combustion engine through vibration and acoustic pressure non-intrusive measurements. Appl. Therm. Eng. 2009, 29, 1707–1713. [Google Scholar] [CrossRef]
  42. Jedliński, Ł.; Caban, J.; Krzywonos, L.; Wierzbicki, S.; Brumerčík, F. Application of vibration signal in the diagnosis of IC engine valve clearance. J. Vibroeng. 2015, 17, 175–187. [Google Scholar]
  43. Mahdisoozani, H.; Mohsenizadeh, M.; Bahiraei, M.; Kasaeian, A.; Daneshvar, A.; Goodarzi, M.; Safaei, M.R. Performance Enhancement of Internal Combustion Engines through Vibration Control: State of the Art and Challenges. Appl. Sci. 2019, 9, 406. [Google Scholar] [CrossRef]
  44. Tamura, M.; Saito, H.; Murata, Y.; Kokubu, K.; Morimoto, S. Misfire detection on internal combustion engines using exhaust gas temperature with low sampling rate. Appl. Therm. Eng. 2011, 31, 4125–4131. [Google Scholar] [CrossRef]
  45. Avan, E.Y.; Mills, R.; Dwyer-Joyce, R. Ultrasonic Imaging of the Piston Ring Oil Film During Operation in a Motored Engine—Towards Oil Film Thickness Measurement. SAE Int. J. Fuels Lubr. 2010, 3, 786–793. [Google Scholar] [CrossRef]
  46. Ranachowski, Z.; Bejger, A. Fault Diagnostics of the Fuel Injection System of a Medium Power Maritime Diesel Engine with Application of Acoustic Signal. Arch. Acoust. 2005, 30, 465–472. [Google Scholar]
  47. Hwang, O.; Lee, M.C.; Weng, W.; Zhang, Y.; Li, Z. Development of novel ultrasonic temperature measurement technology for combustion gas as a potential indicator of combustion instability diagnostics. Appl. Therm. Eng. 2019, 159, 113905. [Google Scholar] [CrossRef]
  48. Förster, F.; Crua, C.; Davy, M.; Ewart, P. Temperature measurements under diesel engine conditions using laser induced grating spectroscopy. Combust. Flame 2019, 199, 249–257. [Google Scholar] [CrossRef]
  49. Chen, J.; Randall, R.B.; Feng, N.; Peeters, B.; Van der Auweraer, H. Automated Diagnostics of Internal Combustion Engines using Vibration Simulation. In Proceedings of the ICSV20, Bangkok, Thailand, 7–11 July 2013. [Google Scholar]
  50. Lima, T.L.; Filho, A.C.L.; Belo, F.A.; Souto, F.V.; Silva, T.C.B.; Mishina, K.V.; Rodrigues, M.C. Noninvasive Methods for Fault Detection and Isolation in Internal Combustion Engines Based on Chaos Analysis. Sensors 2021, 21, 6925. [Google Scholar] [CrossRef] [PubMed]
  51. Rodrigues, N.F.; Brito, A.V.; Ramos, J.G.G.S.; Mishina, K.D.V.; Belo, F.A.; Lima Filho, A.C. Misfire Detection in Automotive Engines Using a Smartphone through Wavelet and Chaos Analysis. Sensors 2022, 22, 5077. [Google Scholar] [CrossRef]
  52. Monieta, J. Diagnosing Cracks in the Injector Nozzles of Marine Internal Combustion Engines during Operation using Vibration Symptoms. Appl. Sci. 2023, 13, 9599. [Google Scholar] [CrossRef]
  53. Wieclawski, K.; Figlus, T.; Mączak, J.; Szczurowski, K. Method of Fuel Injector Diagnosis Based on Analysis of Current Quantities. Sensors 2022, 22, 6735. [Google Scholar] [CrossRef]
  54. Pająk, M.; Kluczyk, M.; Muślewski, Ł.; Lisjak, D.; Kolar, D. Ship Diesel Engine Fault Diagnosis Using Data Science and Machine Learning. Electronics 2023, 12, 3860. [Google Scholar] [CrossRef]
  55. Terwilliger, A.M.; Siegel, J.E. Improving Misfire Fault Diagnosis with Cascading Architectures via Acoustic Vehicle Characterization. Sensors 2022, 22, 7736. [Google Scholar] [CrossRef]
  56. Smart, E.; Grice, N.; Ma, H.; Garrity, D.; Brown, D. One Class Classification Based Anomaly Detection for Marine Engines. In Intelligent Systems: Theory, Research and Innovation in Applications; Jardim-Goncalves, R., Sgurev, V., Jotsov, V., Kacprzyk, J., Eds.; Studies in Computational Intelligence; Springer International Publishing: Cham, Switzerland, 2020; Volume 864, pp. 223–245. ISBN 978-3-030-38703-7. [Google Scholar]
  57. Lo, C.K.; Chen, C.H.; Zhong, R.Y. A review of digital twin in product design and development. Adv. Eng. Inform. 2021, 48, 101297. [Google Scholar] [CrossRef]
  58. Xiong, M.; Wang, H. Digital twin applications in aviation industry: A review. Int. J. Adv. Manuf. Technol. 2022, 121, 5677–5692. [Google Scholar] [CrossRef]
  59. Yu, G.; Wang, Y.; Mao, Z.; Hu, M.; Sugumaran, V.; Wang, Y.K. A digital twin-based decision analysis framework for operation and maintenance of tunnels. Tunn. Undergr. Space Technol. 2021, 116, 104125. [Google Scholar] [CrossRef]
  60. Granacher, J.; Nguyen, T.-V.; Castro-Amoedo, R.; Maréchal, F. Overcoming decision paralysis—A digital twin for decision making in energy system design. Appl. Energy 2022, 306, 117954. [Google Scholar] [CrossRef]
  61. Li, J.; Zhou, G.; Zhang, C. A twin data and knowledge-driven intelligent process planning framework of aviation parts. Int. J. Prod. Res. 2022, 60, 5217–5234. [Google Scholar] [CrossRef]
  62. Wu, Z.; Li, J. A Framework of Dynamic Data Driven Digital Twin for Complex Engineering Products: The Example of Aircraft Engine Health Management. Procedia Manuf. 2021, 55, 139–146. [Google Scholar] [CrossRef]
  63. Jiang, J.; Li, H.; Mao, Z.; Liu, F.; Zhang, J.; Jiang, Z.; Li, H. A digital twin auxiliary approach based on adaptive sparse attention network for diesel engine fault diagnosis. Sci. Rep. 2022, 12, 675. [Google Scholar] [CrossRef]
  64. Stoumpos, S.; Theotokatos, G. A novel methodology for marine dual fuel engines sensors diagnostics and health management. Int. J. Engine Res. 2022, 23, 974–994. [Google Scholar] [CrossRef]
  65. Tsitsilonis, K.-M.; Theotokatos, G.; Patil, C.; Coraddu, A. Health assessment framework of marine engines enabled by digital twins. Int. J. Engine Res. 2023, 24, 3264–3281. [Google Scholar] [CrossRef]
  66. Aghazadeh Ardebili, A.; Ficarella, A.; Longo, A.; Khalil, A.; Khalil, S. Hybrid Turbo-shaft Engine Digital Twining for Autonomous Air-crafts via AI and Synthetic Data Generation 2023. Aerospace 2023, 10, 683. [Google Scholar] [CrossRef]
  67. Liang, J.; Mao, Z.; Liu, F.; Kong, X.; Zhang, J.; Jiang, Z. Multi-sensor signals multi-scale fusion method for fault detection of high-speed and high-power diesel engine under variable operating conditions. Eng. Appl. Artif. Intell. 2023, 126, 106912. [Google Scholar] [CrossRef]
Figure 1. Proknow-C methodology. Adapted from [36,37,38,39].
Figure 1. Proknow-C methodology. Adapted from [36,37,38,39].
Energies 17 06164 g001
Figure 2. Ordinatio methodology. Adapted from [37,40].
Figure 2. Ordinatio methodology. Adapted from [37,40].
Energies 17 06164 g002
Figure 3. Methodology used for the systematic review.
Figure 3. Methodology used for the systematic review.
Energies 17 06164 g003
Figure 4. Search results by database.
Figure 4. Search results by database.
Energies 17 06164 g004
Figure 5. Methodology for filtering.
Figure 5. Methodology for filtering.
Energies 17 06164 g005
Figure 6. History of publications of the BP.
Figure 6. History of publications of the BP.
Energies 17 06164 g006
Figure 7. Demographic distribution of BP publications.
Figure 7. Demographic distribution of BP publications.
Energies 17 06164 g007
Figure 8. History of publications of the BP.
Figure 8. History of publications of the BP.
Energies 17 06164 g008
Figure 9. Distribution graph of bibliographic portfolio themes.
Figure 9. Distribution graph of bibliographic portfolio themes.
Energies 17 06164 g009
Figure 10. Distribution chart of sectors that apply digital twin in bibliographic portfolio.
Figure 10. Distribution chart of sectors that apply digital twin in bibliographic portfolio.
Energies 17 06164 g010
Table 1. Keywords and search results by database.
Table 1. Keywords and search results by database.
Searched CombinationsSCOPUSWoSGoogle Scholar
“Noninvasive Methods” AND “Internal Combustion Engines”111180
“Fault detection” AND “Internal Combustion Engines”222109500
“Noninvasive Methods” AND “Fault detection” AND “Internal Combustion Engines”61450
Total articles per database2391111130
Table 2. Bibliographic portfolio—BP.
Table 2. Bibliographic portfolio—BP.
BP ArticlesRef.Cit.
Diagnosis of internal combustion engine through vibration and acoustic pressure non-intrusive measurements[41]135
Application of the discrete wavelet transform and probabilistic neural networks in IC engine fault diagnostics[33]81
Application of vibration signal in the diagnosis of IC engine valve clearance[42]76
Performance Enhancement of Internal Combustion Engines through Vibration Control: State of the Art and Challenges[43]37
Acoustic Diagnostics Applications in the Study of Technical Condition of Combustion Engine[26]30
Fault Detection of a Flow Control Valve Using Vibration Analysis and Support Vector Machine[35]28
Misfire detection on internal combustion engines using exhaust gas temperature with low sampling rate[44]26
Ultrasonic Imaging of the Piston Ring Oil Film During Operation in a Motored Engine—Towards Oil Film Thickness Measurement[45]25
Fault diagnostics of the fuel injection system of a medium power maritime diesel engine with application of acoustic signal[46]18
Development of novel ultrasonic temperature measurement technology for combustion gas as a potential indicator of combustion instability diagnostics[47]16
Temperature measurements under diesel engine conditions using laser induced grating spectroscopy[48]14
Automated diagnostics of internal combustion engines using vibration simulation[49]8
Noninvasive Methods for Fault Detection and Isolation in Internal Combustion Engines Based on Chaos Analysis[50]6
Misfire Detection in Automotive Engines Using a Smartphone Through Wavelet and Chaos Analysis[51]3
Diagnosing Cracks in the Injector Nozzles of Marine Internal Combustion Engines during Operation Using Vibration Symptoms[52]2
Method of Fuel Injector Diagnosis Based on Analysis of Current Quantities[53]2
Ship Diesel Engine Fault Diagnosis Using Data Science and Machine Learning[54]1
Improving Misfire Fault Diagnosis with Cascading Architectures via Acoustic Vehicle Characterization[55]1
One Class Classification Based Anomaly Detection for Marine Engines[56]1
Table 3. Keywords and results of searches in databases.
Table 3. Keywords and results of searches in databases.
Searched CombinationsSCOPUSWoSGoogle Scholar
(“Digital Twin” OR “Digital-Twin” OR “Gêmeo Digital”) AND (“internal combustion engines” OR “internal combustion engine”)205999
(“Digital Twin” OR “Digital-Twin” OR “Gêmeo Digital”) AND (“internal combustion engines” OR “internal combustion engine”) AND (“Fault detection”)12284
Total articles per database2171282
Table 4. BP for the digital twin.
Table 4. BP for the digital twin.
BP ArticlesRef.Cit.
A review of digital twin in product design and development[57]229
Digital twin applications in aviation industry: A review[58]81
A digital twin-based decision analysis framework for operation and maintenance of tunnels[59]67
Overcoming decision paralysis—A digital twin for decision making in energy system design[60]55
Digital twin-driven optimization of gas exchange system of 2-stroke heavy fuel aircraft engine[14]46
Towards marine dual fuel engines digital twins—integrated modelling of thermodynamic processes and control system functions[12]42
A twin data and knowledge-driven intelligent process planning framework of aviation parts[61]34
A framework of dynamic data driven digital twin for complex engineering products: the example of aircraft engine health management[62]29
A digital twin auxiliary approach based on adaptive sparse attention network for diesel engine fault diagnosis[63]28
A novel methodology for marine dual fuel engines sensors diagnostics and health management[64]24
Health assessment framework of marine engines enabled by digital twins[65]7
Hybrid turbo-shaft engine digital twining for autonomous air-crafts via AI and synthetic data generation. [66]1
Table 5. Technologies of BP processing methods.
Table 5. Technologies of BP processing methods.
BP Ref.MethodAcquisition SystemProcessing SystemMeasurement FrequencyDiagnostic
InvasiveNoninvasiveSoundVibrationTemperatureCurrentCylinder oilDSPIASampleContinuousReal timeMechanical failuresIgnition failuresCO2Injection failuresThickness of the oil film in the cylinder
[41]
[33]
[42]
[43]
[26]
[35]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[67]
Sample: Data are collected at predefined time intervals. Continuous: Data are collected continuously, without interruption, but are not necessarily processed immediately. Real time: Data are collected and processed immediately, with actions or analyses occurring simultaneously with the measurement.
Table 6. Article classification system that uses AI.
Table 6. Article classification system that uses AI.
Method Classification System that Used AIDatasetAcquisition SystemMeasurement FrequencyDiagnostic
BP RefAlgorithmCategorySampling ratePeriodicityType DatasetPublic?SoundVibrationTemperatureCurrentCylinder oilSampleContinuousReal timeMechanical failuresIgnition failuresCO2Injection failuresThickness of the oil film in the cylinder
[33]PNNANN200 samples50% training
50% test
Not specifiedNot
[42]MLPANN------Experimental Not
[35]SVMKernel method4000 samplesNot specifiedExperimental Not
[49]MLP e PNNANN593 samples
101 experimental
492 simulation
100% trained by simulated data
100% trained by experimental data
Experimental and simulation models Not
[50]ANNANN1440 samples60% training
40% test
Experimental Not
[52]ANNANN------Not specifiedNot
[54]SVMKernel method16,384 samples80% training
20% test
Experimental Not
[55]CNNANN286 samples80% training
20% test
Experimental Not
[56]OCSVMANN4.8 kHzNot specifiedExperimental Not
[67]MFRCNNANN51.2 kHzNot specifiedExperimental Not
Sample: Data are collected at predefined time intervals. Continuous: Data are collected continuously, without interruption, but are not necessarily processed immediately. Real time: Data are collected and processed immediately, with actions or analyses occurring simultaneously with the measurement.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Torres, N.N.S.; Lima, J.G.; Maciel, J.N.; Gazziro, M.; Filho, A.C.L.; Souto, C.R.; Salvadori, F.; Ando Junior, O.H. Non-Invasive Techniques for Monitoring and Fault Detection in Internal Combustion Engines: A Systematic Review. Energies 2024, 17, 6164. https://doi.org/10.3390/en17236164

AMA Style

Torres NNS, Lima JG, Maciel JN, Gazziro M, Filho ACL, Souto CR, Salvadori F, Ando Junior OH. Non-Invasive Techniques for Monitoring and Fault Detection in Internal Combustion Engines: A Systematic Review. Energies. 2024; 17(23):6164. https://doi.org/10.3390/en17236164

Chicago/Turabian Style

Torres, Norah Nadia Sánchez, Jorge Gomes Lima, Joylan Nunes Maciel, Mario Gazziro, Abel Cavalcante Lima Filho, Cicero Rocha Souto, Fabiano Salvadori, and Oswaldo Hideo Ando Junior. 2024. "Non-Invasive Techniques for Monitoring and Fault Detection in Internal Combustion Engines: A Systematic Review" Energies 17, no. 23: 6164. https://doi.org/10.3390/en17236164

APA Style

Torres, N. N. S., Lima, J. G., Maciel, J. N., Gazziro, M., Filho, A. C. L., Souto, C. R., Salvadori, F., & Ando Junior, O. H. (2024). Non-Invasive Techniques for Monitoring and Fault Detection in Internal Combustion Engines: A Systematic Review. Energies, 17(23), 6164. https://doi.org/10.3390/en17236164

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop