Murtadha Shukur Mahmood*
| Israa Rafie Shareef
© 2024 The authors. This article is published by IIETA and is licensed under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).
OPEN ACCESS
This survey artісle provides an expose оf the relative literature with a specific focus in conveyor belt systems incorporation оf artificial intelligence (AI). This survey artісle describes the belt condition and its prognostics based on IoT, performance analysis, visualization, and force mailing. The review is based on over 79 articles of peer-reviewed journals published in the last five years of the study and focused on the enhancement of performance and safety of conveyor belt systems for manufacturing, mining, and logistics industries applying advanced AI techniques using DL models. The AI technologies to be investigated are majors of the ML algorithms to be used in the detection of faults and prediction of failures, CV systems to be used in real-time identification of defects on the assets and IoT systems to be used in the collection and processing of data. From the survey, it is seen that the integration of these set of possibilities of AI enhances the competency in the areas of accurate fault detection; superior control and computer based intelligent operation of the material handling than the aspect of monitoring the fan conveyor. Innovation involves some concepts that include the following; belt tear prediction models using neural networks with more than 95 percent certainty for the real time prediction of belt tears, computer vision, for the real time identification of surface issues, IoT that can reduce the system’s unplanned time by at least 30 percent. It also describes the current state of affairs when it comes to data quality problem, explanation of the algorithms used and the procedure of scaling up the already existing systems. Last but not least, it offers key and precise recommendations for the further research on the multiple levels of intelligence in AI systems as well as the Edge AI intelligent decisions, the Reinforcement Learning intelligent control, and AI with other emerging technologies; Digital Twin. Finally, it might be mentioned that, concerning the survey made, it is possible to state how the conveyer belt system may be altered with the competent usage of the AI in various fields for making performance, reliability, as well as security improvements.
AI systems in conveyor belt, CNN models, densenet classification, learning algorithms, visual recognition
Conveyor belts are integral to conveyancing equipment implemented in production lines, supply chain, and mining industries among others. The request of production and streaming of products within manufacturing plants, warehouses and distribution centers rely with these systems. Conveyor belt system improvisation has emerged as а crucial area for concern because of the increased demand by entrepreneurs for high productivity. The conventional methods of operating and managing conveyor belts called for regular monitoring and maintenance schedules, whereby the process always experienced unwanted downtimes and inefficiencies [1].
Incorporating AI into conveyor belt systems can be considered as a successful approach to increase the reliability of the systems as well as their maintenance levels. AI technology ranches like IoT, computer vision, and machine learning are superior technologies for supervising conveyor belts. Thus, the development of intelligent systems that can predict failures, prevent them, and adjust parameters to improve system efficiency as a whole might be attained with the help of these technologies. The integration erases the limitations of the traditional conveyor belt system while embracing the advanced and automated new approach of material handling [2].
There is no doubt that conveyor belts are incredibly valuable in a wide variety of businesses. To the mining industry, conveyors are applied to move valuable materials from depositing locations to processing centers. It is clear that anyone of these disrupts can lead to high financial losses and poor organizational performance. In a similar manner in manufacturing [3], a conveyor belt helps subassemblies transport components from one assembly line to the next in order for assembly processes to occur without stopping. In logistics and distribution, conveyors play numerous roles in conveying and sorting small packages are crucial in ensuring timely delivery of products to the end customer. These systems are critical since the availability, effectiveness and reliability of these systems can positively influence the performances and the profitability of these industries. Figure 1 shows main elements of conveyor system.
Figure 1. Elements of conveyor system [4]
AI technologies have brought significant changes to different industries, and conveyor belt systems are not an exemption in adopting such technologies. The demand for enhanced automation in the course of working, minimization of time between stops, enhancement of production effectiveness has stimulated the formation and evolution of AI technologies. In conveyor belt systems, frictionless data acquisition from installed sensors through of machine learning will help to discover patterns that are likely to cause a failure. Its platforms allow gathering and processing real-time data about the status оf conveyor belt systems and alerting about the necessity of maintenance [5].
Defining structures that are capable of using data meanings for learning and even predicting is referred to as machine learning, and is a part of artificial intelligence. Machine learning algorithms are used to learn conventioneer belt systems’ data with tends in it and to be able to predict potential issues head оf time. These algorithms are particularly beneficial for real-time defect detection and for prognostics because such algorithms can generally always be honed as more data are received in the future. Some examples are, Neural network may be employed to detect small variations in the signals from sensors that could indicate an imminent failure because they are designed to mimic the brain capability for recognizing patterns in a sequence of concrete data [6].
The physical state оf the conveyor belt is checked through cameras and feature extraction on images that are captured using computer vision, another of the key AI applications. Images and videos can be analyzed by computer vision systems to detect exterior imperfections and measure alignment and any other signs of wear and tear. This system can accurately analyse visual data with help of image analysis techniques such as the edge detection, Hough transform and machine learning. This feature is extremely useful in environs such as high-speed production line or deep mining areas where manual inspection is either difficult or impossible [7].
This systematic review aims to identify and consolidate the most recent studies on the use of AI in conveyor belt systems and their effectiveness across industries. Whereas prior reviews have only presented a selection of AI methods or addressed the situation of particular industries only, this survey presents an overview of different AI technologies that can be combined to develop intelligent and connected conveyor monitoring systems. The novel contributions of this review include: Identification of trends and recent developments in AIV technologies through the following: (1) A comprehensive review of over 70 research papers published in the last few years to capture the current state of knowledge in the field; (2) A comparative analysis of the effectiveness of ML, CV, and IoT solutions for identifying faults, predicting maintenance needs, and optimizing operation; (3) A review of current issues and shortcomings associated with the use of AI in conveyor.
2.1 Traditional conveyor systems
Old-fashioned conveyors have been used in industries for a long time but present themselves with some disadvantages for instance, maintenance that may take a long time leading to halt on operations and thus, increased expenses. This section elaborates the problems seen in traditional systems for instance; the task of predicting when maintenance is due and how to reduce failures [8].
2.2 Integration of artificial intelligence
This section explains the way in which technologies like, machine learning, computer vision, IoT can improve the conveyor system. It describes the overall application of such things as artificial intelligence's role in enhancing the reliability, performance, and maintenance of the existing systems [9].
2.3 IoT platforms
Internet of Things is among the foundational aspects of applying Artificial Intelligence to conveyors in that it makes it possible to gather real-time data and generate insights. This sub section looks at how IoT sensors monitor different parameters of systems for temperature, vibration, speed among others for system health forecast and system efficiency [10].
2.4 Machine learning for fault detection
For fault detection and for the knowledge of when it is time for a machine to be serviced, machine learning is very important. This section deepens the topic of employing the neural networks and support vector machines (SVM) in the examination of the senator data for the aim of finding the recurrent failures and then to notify the operators of the failures on-line [11].
2.5 Computer vision for monitoring
Automated technologies in the field of computer vision improve the use of pictures and videos locked to the articulation for detecting surface irregularities and mismatches. This section describes edges and the Hough Transform used in analysis of features and wear and tear on the belts on conveyors [12].
2.6 Predictive maintenance strategies
Corporate AI analyses data both historical and real-time data to be able to predict when maintenance is likely going to be required thus avoiding plan breakdowns. This sub-section describes how the AI models such as the dynamic mathematical models help in the prediction of failures and scheduling of the maintenance [13].
2.7 System optimization/improvement
AI technology enhances works parameters like the belt speed, load's arrangement on it, etc., that minimizes energy consumption. This section looks at the extent to which AI has influenced the overall systems performance for instance by enhancing reliability of the system and productivity of those implementing the system [14].
As a review paper this paper will compile data from at least, 70 research articles related to AI based conveyor belt system. The material of these articles deals with Fault Detection, Predictive Maintenance, optimization, and System Performance enhancement. The approach that has been followed entails the following: Content analysis of the findings has been done to categorize the outcome into major themes; Summarization of methodologies utilized in the choice of the studies; and most importantly, the discussion on the significance of the outcome of the studies.
3.1 Research design
The study method that is employed for this research is systematic literature review whereby the purpose of identifying publications related to the integration of AI technologies in conveyor belt systems for both prediction and system improvement. This process involves a critical analysis of the current literature in the academic and business fields.
3.2 Search strategy
Having a pool of diverse papers was obtained through the use of the following electronic databases namely IEEE Xplore, ScienceDirect, SpringerLink and Google Scholar that was searched systematically. The search terms used were:
"AI in conveyor systems"
"Predictive maintenance using AI"
"IoT in industrial automation"
For instance, using of machine learning techniques in fault identification and in future forecasting.
"Computer vision in manufacturing"
These terms were chosen to encompass all the general areas of AI technologies used in conveyor belt systems.
3.3 Inclusive and exclusive criteria
The papers were selected based on the following inclusion criteria:
Relevance: The research needs to be centered toward applying the existing AI technologies like, machine learning, IoT and computer vision in conveyor belt systems.
Publication Date: Recent papers of the last decade (2013-2023) that allow focusing on technologies and trends of the present days.
Language: Data from only those papers which were written in English language only have been included.
Exclusion criteria included: Non-peer-reviewed Sources: Any papers that were non-refereed in nature were hence not included in the literature review in order to have a good quality literature.
Irrelevant Topics: Only research works that focused on conveyor systems and/or the use of predictive maintenance methodologies were excluded from this analysis.
3.4 Paper selection process
The preliminary search produced 250 papers. title and abstract screening were done to leave only 120 papers. These papers were further judged for the quality and relevance according to the inclusion and exclusion criteria and that led to 50 papers for full text consideration.
3.5 Data extraction and data synthesis
Relevant data such as research goals and objectives, methods of data collection and analysis, AI technology employed and the outcomes concluded were also gathered from the chosen papers. The results were structured around concepts that relate to various AI uses in conveyors, including maintenance prediction and conveyor arrangement.
The synthesis process involved:
Thematic Analysis: Proactive comparison of papers selected: analysis of general trends and understanding of the main messages.
Comparative Analysis: Drawing a comparison between different AI technologies and an understanding of which are most advantageous for improvement of the conveyor belt systems.
Paper Hitherto, ML has been posited as a beneficial approach in conveyor belt systems particularly as it relates to fault identification and prediction [15]. But it is also true that through the machine learning analysis can be made of data gathered through the different sensors of the system and it would alert the status or future issues with the system at any given time. Neural networks operating systems such as SVM and decision tree are known to be applied in conveyor belt systems. These are helpful in big datasets and ascertainment of substructures which may well be latent and not discerned through other nondescript methods [16].
Figure 2. Structure оf neural network for condition recognition [14]
Artificially designed NEURAL NETWORKS that replicate the brain's work capacity are very effective in jobs that may include pattern recognition and or forecasting. They can also gaze at data acquired from sensors placed in conveyer belt systems and detect such things like odd vibrations or high temperatures which are likely to point to system failure [17]. SVMs are used for both classification and regression issues, this involves the determination of the hyperplane that provides optimal margin of separation between normal operation data and other faulty states that may be present within data space [18]. It can be used as a kind of algorithm in classification and regression and besides, it is not complex to comprehend. It can be useful in deciding the exact conditions under which a fault is more likely to occur so as to allow proper maintenance to be done on it right [19]. For instance, the structure of a neural network for condition recognition, such as the one presented in this piece of work, is illustrated in Figure 2.
According to the study performed by Wang et al. [20] certification was made that all belt deviations and other faults could be diagnosed by the use of machine learning models. On this basis, through training of technique models оf historical data it culminated in the report that they developed high accuracy in predicting the failures before they occur so as to anticipate intervention for maintenance.
Other computer vision technologies have also enriched the real time monitoring as well as the maintenance of conveyor belt systems. All these technologies permanently use cameras and image analysis algorithms to monitor the conditions of conveyor belts [21]. Contouring methods such as edge detection can help in enhancing crack, tear or any other surface distortion whilst Hough transform helps in making out the circle or any line in the image depending on misalignment or uneven wear on the belt [22]. This knowledge is supported by previous studies, showing that hone machine learning algorithms, such as convolutional neural network (CNNs), exceeding the capability to be trained to distinguish pointed types of defects. For example, a dynamic model of belt conveyors including a vibrational implement censor and a tension censor is shown in Figure 3.
Figure 3. Dynamic model оf belt conveyor: (a) Vibration sensor; (b) Tension sensor [21]
Real time analyses are indeed a distinctive feature of computer vision systems because it averts the operators about the drawback on the spot enabling them to correct it. As shown by Wu et al. [12], this paper also highlights that belt deviation faults can also be diagnosed using the machine vision technology with admirable accuracy and stability. The monitoring of the conveyor belt condition together with constant monitoring for fault development allow the constant observation of the conditions affecting the conveyor belt and reduces the likelihood of serious failures while at the same time extending the life expectancy of the systems.
The Internet of Things (IoT) does have part to play in the current conveyor belt system as it improves on the functions of the system in a view of improving on therefore belt as a data gathering tool in real-time applications. IoT platforms facilitate the devices and sensors to gather information regarding various parameters of operations that are conveyed to the main systems for further evaluation. Sensor’s equipment is good at offering extensive monitoring and better reliability as pointed out in the study [23]. For instance, we have the dynamic model that shows one or many sensors, and device involved in the control of a conveyor belt as illustrated in Figure 4.
Figure 4. Integration оf various sensors and devices for monitoring a conveyor belt [1]
IoT platforms act as the interface through which they receive data collected by IoT sensors and process and analyze it with the help of AI algorithms. These platforms are capable оf providing real-time information concerning the status оf the conveyor belt system. Zhang et al. [14] reported about the integration of AI models with IoT for accurate fault diagnosis and timely predictive maintenance, monitoring and command distant and enhanced flexibility and productivity.
An all-inclusive AI based Conveyor belt system that can tackle the issues concerning the traditional systems is developed with the aid of machine learning, computer vision, and IoT. They include; ongoing assessment, real-time defect identification, prediction of failure, and process parameter optimization [21]. For instance, location of the conveyor belt between the two adjacent idlers is shown in Figure 5 where multiple monitoring locations are labeled.
Machine learning in conveyor belt system HMIs has also revolutionized the way predictive maintenance is carried using AI models to develop models for estimating the remaining useful life of parts. These models assess numerical data from the sensors in addition to the records of prior maintenance and operations logs in order to forecast the future potential of the component failure successfully. As mentioned above, these models are advantageous because companies can identify maintenance requirements before any breakdown hence avoiding frequent conveyor belt failure, determine the right time for conveyor belt maintenance as well as increase the service duration of each conveyor belt part [24].
Figure 5. Structure оf the conveyor belt between two adjacent idlers [21]
Present dons fault detection techniques using Moschino learning have been a great effective in fault detection in conveyor belt systems. All these methods employ historical information to produce models that can identify pates suggesting fault. Zhang et al. [14] introduced a fault diagnosis method with IoT platforms combined with the LightGBM model successfully in identifying multiple faults, including belt slippage, misalignment, as well as wear. The application оf complex algorithms improves accuracy and efficiency in the fault diagnosis programs.
Figure 6. Flow chart оf experimental work and ANN [1]
The range of usage of CVS is found to be very significant in the determination of the surface d frost observed on the convection belts. For instance, Wu et al. [25] designed a machine vision-based system for diagnosing belt deviation faults in coal mines using image processing toolbox, such as adaptive Hezbollah reflex ion and edge detection, and these systems continuously monitor for damages such as tears, cuts and abrasions on conveyor belts with possibility of likely repair thereby affecting the operations of the conveyors. An example flow chart of experimental work, as well as the application of artificial neural networks (ANN) In fault detection is illustrated in the following Figure 6: An example flow chart of experimental work, as well as the application of artificial neural networks (ANN) In fault detection is illustrated in Figure 6.
Hence there have been tremendous add-ons due to integration of machine learning, IoT and, computer vision in conveyor belt systems to enhance fault detection and precise working of conveyor belt systems in terms of maintenance [26]. These technologies also enhance an ability to observe the status of conveyor systems as they operate; to be able to identify that there are faults during the early stages of operations; and to be able to schedule proper maintenance time for the conveyor belt systems, which results in higher reliability, optimum performance, and durability for the belt conveyor systems [27].
Actually, these days Condition Monitoring as well as Conveyor Belt Management and Maintenance are governed by the advanced ML algorithms [28]. Many of these algorithms are refined and excellent in conjuring elaborate solutions in data and patterns for problems related to detection of faults and efficiency in maintenance and systems improvements. Therefore, it is necessary to develop the separate ML methodologies and consider how each of them can be used to identify the problems associated with conveyor belts and how they can be resolved [29].
5.1 Neural networks
Recently, neural networks have been applied in monitoring of conveyor belt due to their capability in pattern recognition of nonlinearity [30]. Based on specifics of the used architectures, it is possible to point to Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) as effective ones in this particular domain.
Convolutional Neural Networks (CNNs): CNNs are quite significant for image-based monitoring operations. For example, CNNs have been used to analyze image of conveyors to identify possible defects on the belt surface, possible tears, and misalignment. In one of the implementations, a CNN was trained with the help of a dataset of labeled conveyor belt images to detect the defects with an accuracy of 95%. There were several convolutional layers which are followed by the pooling layers and the fully connected layers, which provided the model with hierarchical features [31].
Recurrent Neural Networks (RNNs): RNNs are good for time series prediction and belt wear and tear has been predicted using RNNs [32]. In a particular study, for the purpose of predicting the conveyor belt failures through the analysis of the past data read by sensors, including temperature and vibrations, an RNN model with the LSTM units was designed. In the case of the LSTM model used in the prediction, the achieved accuracy was 92%, while it was possible to predict possible failures and provide early warning for performing necessary maintenance operations [22].
5.2 Support vector machines (SVM)
Support Vector Machines are used in fault diagnosis or fault identification of conveyors. They apply most, if not all, of the time for electric measurements intent on differentiating between normal and faulty situations [33].
Specific Implementations: SVM was used in this study to classify motor current of the conveyor belts as normal or abnormal. The current features which were employed were the current amplitude, current frequency and the levels of the current’s harmonic distortion. In the present work, the first SVM model yielded 89% classification accuracy for the given electrical faults in motors [34].
Feature Selection: SVMs highly rely on feature selection to give optimal results on the classifier. And in conveyor monitoring one can use belt speed, vibration frequency, and temperature as some of the parameters. Thus, for these features, an SVM model employing an RBF kernel was applied, which could effectively detect mechanical fault types, such as pulley misalignment and roller defect [34].
5.3 Decision trees
Decision trees can enable the diagnosis of conveyor belt issues from a point of view of the model in a clear and easily understandable manner. They are used to develop the decision rules on the input features preceding various fault conditions.
Specific Implementations: Different diagnostic systems for the condition of conveyor belts have been created based on decision tree algorithms. One example of using ML is when a decision tree model was learned from data that the company gathered from acoustic sensors that monitor the sound signals generated by the conveyor parts. The coefficient of accuracy of the model was 87% in terms of detecting such faults as belt slippage and roller bearing failure [35].
Feature Importance: They also provide information on the relative significance of the features; thus, their application sheds light on which parameters should be closely monitored. For instance, a study established that the magnitude of vibration and belt tension were some of the critical predictors of failure. Due to the decision tree construct, one can readily identify elements of systems that require preventive work, to be performed by the maintenance teams [20].
5.4 Performance comparison
Comparison of Techniques: In the case of techniques, CNNs had the most accuracy for image-based defect identification purposes, while LSTMs were more effective in time series. Hence, SVMs and decision trees offered good and quite stable solutions for fault classification, while the former was more accurate for linearly separable data, and the latter, being explainable [36].
Integration and Hybrid Approaches: More advanced work has focused on the integration of different AI methods with an aim of improving the results obtained. For instance, incorporating CNNs with LSTMs enables spatial and temporal features analysis which helps increase the defect detection and failure prediction accuracy [37].
5.5 LightGBM
LightGBM is an effective gradient boosting decision tree framework that is particularly fast and scalable when dealing with large datasets. For condition monitoring in conveyor belt systems, LightGBM is used in fault diagnosis and failure prediction where ifs perform an excellent job in the detection of different types of faults such as the belt slipping, belt misalignment, and belt wear [38].
The primary strong suit of LightGBM, therefore, is its suitability for dealing with heterogeneous features and their nonlinear interactions. This makes it appropriate for combining various sensor data flows and maintenance history to detect failure patterns [39]. LightGBM utilizes ensemble learning to improve the accuracy and credibility of the fault identification systems; more importantly it helps maintenance team in their decision making about interventions for improving system efficiencies [39].
Using these algorithms, computer vision, and IoT, conveyor belt systems are enhanced to the extent where they can work with minimal defects and maximum efficiency. Such algorithms help to keep track of continuous states of the system while preventing faults in real-time and also determining the right time for maintenance hence cutting down on the overall cost of operations and unnecessary downtime [40].
The integration of computer vision and Internet of Things (IoT) has therefore revolutionized the way conveyor belt systems are supervised and maintained. This synchronized integration not only improves the performance of the monitoring aspects but also the element of predictive maintenance, which greatly improves the effectiveness and reliability of conveyor belt systems [41].
Computer vision technologies have become crucial components in conveyor belt systems, as these technologies allow assess the state оf belts with high accuracy and at the fast rate [42]. These systems use high-resolution cameras and complex image analysis algorithms to investigate the surface of conveyor belts with high precision, even the slightest imperfections are not overlooked [43].
At the same time, IoT platforms have an equally important role in introducing conveyor belts to the age of intelligent manufacturing. By incorporating a variety of sensors and devices into the conveyor framework, IoT platforms enable real-time monitoring plans for many operational characteristics [44]. These sensors include temperature, vibration, speed, and tension sensors that offer important information on the dynamic working оf the conveyor belt system [45].
On the integration оf computer vision and IoT technologies goes beyond simple data harvesting, these connected systems can detect minor variances in operating parameters and act as alarms for emerging faults or variations [46].
Promising studies conducted by Li and Zhang [34] outlined the possibility of applying computer vision and IoT conceptions in conveyor belt systems for fault detection and further maintenance. it allowed them to reach unparalleled levels of accuracy in terms of defect detection and maintenance prediction. In the same vein, pioneered research to explicate the evolutional effect оf IoT platforms integrated with computer vision systems in the management of conveyor belts through real-time monitoring and controlling, thus promoting improved reliability and operational reliability.
This is why the development of conveyor belt system monitoring and maintenance is almost ready to start using the integration of IoT and computer vision technologies. Such integrated systems used for performing preventive maintenance interventions, reduce time for conveyor belt downtime and increase the operating duration of all mechanized parts of the conveyor belt systems through real time performance and condition monitoring [47].
From the following system, computer vision systems are very vital when it comes to identifying surface imperfection on the conveyor belts. Wu et al. [25] proposed a machine vision-based approached to predicting the belt deviation fault has been detected in the coal mines, and high dynamic range imaging and edge detection algorithms used for this purpose. It is always on the lookout for developing problems like tear, cut, or abrasion on conveyer belts so as to allow corrective actions to be instituted early so as not to favor their worsening and thus prolong time off.
The collected data streams are processed through computer vision algorithms with cameras mounted along the conveyor belt to monitor for any complex patterns and flaws. Through a profound learning approach, such systems are capable of differentiating between typical abrasion and real concerns such as severe problems that need to be addressed immediately [48].
Predictive Maintenance
Predictive maintenance is an effective and more strategic way of carrying out maintenance on equipment because it seeks to prevent the failure of the equipment by giving a prior notice [49].
Deadly Sins That Lead to Predictive Maintenance Failure
Different machine learning algorithms have been used to implement and enhance the concept оf predictive maintenance in conveyor belt systems with the help of AI models for estimating the useful life оf components. Oleh and Maksym [50] have employed the modeling approach in order to determine the transient belt stretch during full load starting and schedule the maintenance to predict or avoid belt failures. These models use the raw data from the sensors, past maintenance records, and system logs to estimate the risk level оf an element breakdown reliably.
Regression analysis, artificial neural networks, there are various algorithms that these predictive maintenance models apply, which assist to decrease the unpredicted down time, identify the right time for maintenance of the conveyor belt, and increase the durability оf the belt components [51].
IoT integrates centralized platforms that capture various operation parameters such as temperature, vibration, and speed that relay critical information about the states of the conveyor belt components. Maintenance teams can use this data to identify the need for preventive measures against downtime through AI algorithm’s analysis of such data and adequate identification оf potential breakdowns [52].
Figure 7 shows the macro flow process that is resent in this CRITICAL area оf conveyor belt management and improvement [53]. Conveyer belt surface defects are important and detected primarily by the computer vision systems as described in the preceding text. Later Wu et al. [25] proposed a machine vision-based technique that used image processing and analysis via enhanced imaging procedures for diagnosing belt deviation fault in the coal mines. These systems employ adaptive high dynamic range imaging and edge detection algorithms in order to inspect the conveyor belts for such edges as tears, cuts or abrasion on a permanent basis.
In Figure 7, the training and validation process of DNN models which serves as the core mechanism for precise defect recognition is illustrated.
Figure 7. Macro flow for training and validation оf DNN models [6]
It is crucial to note that improving the main operational factors that define the effectiveness of the conveyor belt systems, together with energy consumption and maintenance costs, is a vital task for improving the efficiency and performance rates of such facilities [54].
AI-Based Optimization Techniques
Their capability of making real-time decisions as well as processing huge amounts of data allows AI algorithms for controlling the operational parameters including belt speed, tension, and load distribution. Kirkire et al. [30] provided a synthesis and evaluation of an inclined belt conveyor system for coal transportation enhancing feasible material transport efficiency by varying the geometric features of width, inclination, etc., under an AI consideration. These techniques include: In this process, there is analysis and dynamic optimization of how these operations work with an aim of making them better, use lesser energy and be more productive.
Machine learning-based approaches in business optimization involves the use of artificial intelligence algorithms to examine operation data to discover new ways of improvement. Through constant feedback and control of the operation conditions, these methods allow for the achievement of the desired flow rate of the conveyor belt while minimizing the proportions of power usage and maintenance needed [55].
Adaptive Control Systems
Adaptive control approach employs AI to update control parameters in an organized form based оf the present working conditions in order to increase productivity. In the work by Zhou et al. [56] included an adaptive automation technique for tensioning device of belt conveyors, which used hydraulic drive to control the belt tension and mitigate abrasive wear. These systems help to respond to changes in the operational conditions by varying some parameters in such a way that reuse will be maximized, availability and reliability will be increased and greatly reduced demands оf maintenance.
Self-regulating control networks are always assessing matters of operation, including but not limited to loads and the environment, maintenance cost [54].
System Performance Improvements
Variables that were taken to optimize and redesign the conveyor belt systems include the efficiency, reliability and overall safety in operation through the application of innovative technologies [28].
Energy Efficiency
Energy efficiency is therefore considered an essential factor when managing conveyer belt since it affects the costs of running the system as well as the environmental conservation. Marijić et al. [57] devoted to the investigation of the artificial friction coefficient of the belt conveyors as well as the influence of this factor to energy consumption by presenting the instructions concerning the belt conveyor resistance to motion. Machine learning algorithms work Autonomously with Data to discover Optimal Trends and adjust processes to use less energy while Node List performance is released or enhanced.
Using parameters such as belt speed, tension and load distribution, these techniques employ the use of machine learning in studying the working data of mechanisms and rolling mills to achieve energy saving results that enhance operational efficiency and thus the reduction of energy output impacts [35].
Reliability and Safety
It is imperative to make sure that conveyor belt systems are also reliable and safe, so that there are no avoidable accidents, and that the system does not break down too often. Additionally, it has been highlighted the usage of magnetic markers for diagnostics, increasing the effectiveness of maintenance activities and elimination of possible failures. The Artificial Intelligence monitoring systems constantly feed data from sensors to identify trends anԁ foresee failures so as to give early signal Fu accidents and frequent breakdowns, which Is helpful in enhancing safety and dependability оf conveyor belt systems.
Figure 8. Structure оf CMMS-class Dia Manager system for conveyors [15]
In Figure 8, the structure оf the CMMS-class Dia Manager system is illustrated and the integration оf high-tech diagnostic tools for more effective maintenances are presented. One of the parameters discussed earlier in regard to the diagnostics is the implementation оf magnetic markers. Such markers help in performing a preventive measure, which means that any irregularity that may be identified as having the potential of leading to a system failure can be detected before it happens, and this leads to better systems' dependability and safety.
Figure 8 depicts the CMMS-class Dia Manager system as a key node for the diagnosis of conveyor belt conditions and as a hub for executing the maintenance activities. This framework also enables monitoring, recognition of irregularities, and prediction оf failures in real-time through AI-assisted analysis of sensor data. In this respect, maintenance teams may be able to schedule jobs and resources more efficiently through proactive measures, thus increasing the reliability and safety of conveyor belt systems [53].
Introducing the artificial intelligence (АІ) in the conveyor belt systems, including for identifying longitudinal tears is the innovative practice which improves the effectiveness оf the maintenance processes. This section discusses the AI approaches used in recent works, particularly on audio-based detection systems and use of Dense Net neural networks. They do not only enhance the identification of faults but also reduce time off-line and expenses on maintenance [1].
8.1 Sound based detection using dense net
Sound-based detection utilizes vibrations made by the conveyor belt during its operation. The method starts with acquiring sound signals which are then used to extract feature using the MFCC and ends with classification using an improved Dense Net model [2].
8.2 Sound signal acquisition
Sound signals are captured through microphones positioned near the conveyor belt. These microphones sample at a frequency оf 48kHz, ensuring high-resolution data collection. Each sound segment spans approximately 2 seconds to encompass various operational states, including different load conditions and belt speeds. To enhance the dataset's robustness, techniques such as time extension, pitch alteration, and background noise incorporation are applied, resulting in a comprehensive and diverse dataset for training and evaluation [3].
8.3 Feature extraction with MFCC
MFCCs are critical for transforming time-domain sound signals into frequency-domain features that highlight the unique characteristics оf tearing sounds. The MFCC extraction process involves several steps [6]:
(1). **Pre-emphasis**: A high-pass filter enhances high-frequency components, which are more indicative оf tears, improving the signal-to-noise ratio.
(2). **Framing**: The continuous sound signal is divided into overlapping frames, typically оf 20-30 milliseconds, to capture short-term spectral features.
(3). **Windowing**: Each frame is windowed, often with a Hamming window, to reduce spectral leakage and focus оf the signal's central part.
(4). **Fast Fourier Transform (FFT)**: The windowed frames are transformed from the time domain to the frequency domain using FFT, yielding a spectral representation.
(5). **Mel Filter Bank**: The FFT spectrum is passed through a Mel filter bank, which mimics the human ear's perception, emphasizing perceptible frequency bands.
(6). **Logarithm and Discrete Cosine Transform (DCT)**: The logarithm оf the Mel-filtered spectrum compresses the dynamic range, followed by a DCT to obtain the MFCCs.
8.4 Dense net classification
The Densen model is a type оf convolutional neural network (CNN) that ensures dense connections between layers, where each layer receives inputs from all preceding layers. This dense connectivity promotes feature reuse and mitigates the vanishing gradient problem, which is crucial for deep network training [12].
To classify the extracted MFCC features, the Dense Net model is enhanced with [20].
Letting the networks emerge deeper and their width larger allows them to detect more intricate sound features. It really goes a long way in differentiating what constitutes normal sounds from those that are indicative of tears.
There are two techniques known as batch normalization and dropout which makes the model, better overall and does not over-fit the training data. The first one keeps the inputs fixed and the other randomly shuts down some of the neural units during training phase. This serves to urge the network to learn at the higher-level features.
There is an element referred to as residual connections that enable information to pass through the network with ease. This allows constructing the networks that are deeper but perform nearly to the same level of efficiency.
In the process of particular classification, in order to segregate things, we need a special formula from mathematics known as cross entropy loss. This tells us how wrong our predictions are regarding labels or how correct we are.
AI methodologies, preferential listening for sound as well as utilizing more complex models like Dense Net, enhance detecting conveyor belt tears even more effective. If we utilize the MFCC for features, we get to learn small details in the sound, which traditional techniques would not pick up on; What is more, how Dense Net is constructed means we get features explored in-depth, hence our detection is very accurate [58].
Some people have been contrasted when various methods were performed, the authors of the above-said algorithms claim these AI approaches are superior, which in several cases get things right for more than 95% of the time. That is why they stand as very accurate and in instances whereby there may be new information they will always adapt. Table 1 shows analysis of research studies and methods applied in conveyor systems.
Table 1. Kye research on conveyor systems and AI
|
Researches |
Year |
Methods Used |
|
Shareef and Hussein [1] |
2021 |
artificial neural network, speed control, power saving |
|
Soares et al. [2] |
2023 |
machine learning techniques, fault diagnosis |
|
Zhukovsky et al. [3] |
2023 |
neuro-control, belt tension adjustment |
|
Orozco et al. [4] |
2024 |
low-cost learning factory, ai teaching |
|
Rothong et al. [5] |
2023 |
PLC-integrated object detection |
|
Klippel et al. [6] |
2022 |
embedded edge AI, longitudinal rip detection, industrial mining environment |
|
Olchówka et al. [18] |
2021 |
statistical analysis, neural network |
|
Stefaniak et al. [19] |
2017 |
data fusion, advanced analytics |
|
Wu et al. [25] |
2023 |
diagnosis and localization system, machine vision |
|
Druzhinina et al. [21] |
2021 |
AI methods, control system modeling |
|
Hao et al. [22] |
2024 |
machine vision, edge detection |
|
Jiang et al. [59] |
2024 |
control system design |
|
Yuan et al. [60] |
2024 |
image processing, detection method |
|
Capelli et al. [23] |
2023 |
AI ethics, clinical surgery |
|
Sun et al. [24] |
2023 |
transverse vibration modelling, tension analysis |
|
Prenner et al. [11] |
2023 |
monitoring system, machine vision |
|
Wu et al. [12] |
2023 |
fault detection technology, machine vision |
In other words, it can be stated that utilization of the applied AI methods, especially sound recognition and Dense Net-based neural networks are the key to maintaining healthy and functional conveyor systems. These methods make this fault detection more accurate and efficient; by doing so, it enables resources in operations to be evenly spent for optimal performance. This truly illustrates the potential of AI in different industries and specifically for big industries.
When applying AI in conveyor belt systems, there are some obstacles that must be met in order to advance the opportunity of conveyor belt systems with AI. It also explicates the future modifications of AI in conveyor belt systems from these challenges.
The first main difficulty can be distinguished that relates to the data quality and availability. To be able to function well, AI algorithms have to be fed with massive amounts of good quality data. Nonetheless, collecting such data from conveyor belt operations is challenging, mainly because the sensors' accuracies are inconsistent, data becomes noisy, and faults are rare occurrences. Addressing these challenges calls for unique solutions with regards to data collection, preprocessing, as well as augmentation to ensure that the AI is reliable and has a wide applicability.
Another big problem is the ability to interpret an AI model a d the ability of an AI model to explain itself. To ensure that the conveyor belt systems are safe to use, different decision-making processes must be clear. However, some of the advanced AI algorithms, like deep neural networks, are quite opaque, and that is why it remains hard to comprehend what they choose. Subsequent research should therefore be directed to creating AI methods that afford explanations of model behaviors that will rebuild trust and confidence among the various stakeholders.
Another issue that complicates them is scalability and adaptability, particularly in large-scale industrial building environments. The application of AI models to various conveyor belt systems and conditions entails efficient transfer learning and very effective algorithms. In addition, the integration with the current infrastructure requires interaction standards and the modular design patterns for creating scalable and transformable systems.
Nevertheless, the growth prospects of AI implementation in conveyor belt systems appear to be quite favorable. A multitude of opportunities for innovation and progression is present and giving ways to the evolution of AI in this regard.
One potential area focus is the synergy between edge computing and AI techniques for real-time decision making at the conveyor belt location. Smart devices with some computation power integrated can perform AI models right on the device, thus lowering latency and bandwidth demands while allowing for quick response to real-time operating environments.
Reinforcement learning and autonomous control are promising domains that can be applied to conveyor belt systems to be self-optimizing. The AI approaches may thus add adaptability and efficiency by enabling conveyor belts to set itself according to various conditions.
Furthermore, together with the further development of new technologies like 5G, digital twins, and blockchain, integration of AI into conveyor belt operations may contribute to increasing the controllability of processes, improving their transparency, traceability, and accountability. These benefits include improving supply chain management and logistics, minimizing wastage in organizations and enhancing sustainability measures through use of AI-activated analytics and monitoring tools.
The application of AI in conveyor belt systems is one of the modern approaches to industrial automation that enhances the simultaneous, advanced conveyor belt operations. Based on the usage of neural networks, support vector machines, as well as decision trees, organizations can improve the control processes, maintenance predictability, and optimize the conveyors. With these AI techniques, one is able to real time monitor the defects, predict the failures and time to carry out the maintenance hence, reducing the downtime as well as operating cost.
CNNs have been found to be accurate with visual disorders on conveyor belts, while RNNs especially the LSTM have been useful for period prediction of damage frequency using the data acquired. Many classifiers such as Support Vector Machines (SVMs) and decision trees have delivered reliable solutions to fault classification and diagnosis and while SVMs are superior in data nonlinearities, decision trees have advantage of explaining the decision-making.
For the researchers, looking at such improvements should emphasize the need to strive more in creating and advancing AI models and better integrated approaches to optimize structures' performance further. This knowledge can be used to apply Artificial Intelligence in conveyor systems to increase productivity, performance, and resource allocation for the practitioners' advantage. Introducing AI to conveyor belts will eventually enhance the reliability of industries and forthcoming ages of creative advances.
[1] Shareef, I.R., Hussein, H.K. (2021). Implementation of artificial neural network to achieve speed control and power saving of a belt conveyor system. Eastern-European Journal of Enterprise Technologies, 2(2): 110. https://doi.org/10.15587/1729-4061.2021.224137
[2] Soares, J.L.L., Costa, T.B., de Moura, L.S., Sousa, W.D.S., Mesquita, A.L.A., Mesquita, A.L.A., Braga, D.D.S. (2023). Comparison of machine learning techniques for fault diagnosis in belt conveyor idlers. In 27th ABCM International Congress of Mechanical Engineering, Florianópolis, Brazil.
[3] Zhukovsky, A., Zhukovsky, V., Tkachov, V., Kravets, G. (2023). Neuro-control adjustment оf conveyor belt tension. Collection оf Research Papers оf the National Mining University, 73: 154-164. https://doi.org/10.33271/crpnmu/73.154
[4] Orozco, E., Cárdenas, P.C., López, J.A., Rodriguez, C.K. (2024). Low-cost desktop learning factory to support the teaching of artificial intelligence. HardwareX, 18: e00528. https://doi.org/10.1016/j.ohx.2024.e00528
[5] Rothong, N., Chinakunwiphat, P., Chainoi, S., Butsanlee, B. (2023). Design of PLC-Integrated object detection for categorizing conveyor. In 2023 Research, Invention, and Innovation Congress: Innovative Electricals and Electronics (RI2C), Bangkok, Thailand, pp. 130-134. https://doi.org/10.1109/RI2C60382.2023.10356010
[6] Klippel, E., Oliveira, R.A.R., Maslov, D., Bianchi, A.G.C., Delabrida, S.E., Garrocho, C.T.B. (2022). Embedded edge artificial intelligence for longitudinal rip detection in conveyor belt applied at the industrial mining environment. SN Computer Science, 3(4): 280. https://doi.org/10.1007/s42979-022-01169-y
[7] Zhu, S., Zhang, S. (2023). Research on measuring method of conveying capacity of mining belt conveyor. Journal of Physics: Conference Series, 2437(1): 012068. https://doi.org/10.1088/1742-6596/2437/1/012068
[8] Kirjanów-Błażej, A., Rzeszowska, A. (2021). Conveyor belt damage detection with the use of a two-layer neural network. Applied Sciences, 11(12): 5480. https://doi.org/10.3390/app11125480
[9] Park, H.M., Kang, B., Van Messem, A., De Neve, W. (2021). 3-D deep learning-based item classification for belt conveyors targeting packaging and logistics. In International Conference on Pattern Recognition. Cham: Springer International Publishing. Springer, Cham, pp. 578-591. https://doi.org/10.1007/978-3-030-68799-1_42
[10] Prayuktha, B., Vishali, M., Alessandro, D., Rodolfo, G. (2022). Quick quality analysis on cereals, pulses and grains using artificial intelligence. In International Conference on Image Analysis and Processing. Cham: Springer International Publishing. Springer, Cham, pp. 372-383. https://doi.org/10.1007/978-3-031-13324-4_32
[11] Prenner, M., Müller, J., Denzel, M. (2023). Monitoring system for belt conveyors. In Proceedings of ICBMH2023-The 14th International Conference on Bulk Materials Storage, Handling and Transportation. https://doi.org/10.34901/mul.pub.2023.153
[12] Wu, X., Wang, C., Tian, Z., Huang, X., Wang, Q. (2023). Research on belt deviation fault detection technology of belt conveyors based on machine vision. Machines, 11(12): 1039. https://doi.org/10.3390/machines11121039
[13] Shamil, Y., Elvira, K., Alexandr, I. (2024). Development оf a control system for the tensioning device оf belt conveyors. Trudy Universiteta, 1. https://doi.org/10.52209/1609-1825_2024_1_295
[14] Zhang, Y., Yuan, C., Zhang, M., Zhang, D., Wang, Z., Wang, Y. (2024). Study on the arrangement of roller sets in the transition segment of the pipe belt conveyor. International Journal of Precision Engineering and Manufacturing, 1-10. https://doi.org/10.1007/s12541-024-01006-7
[15] Gebler, O.F., Hicks, B., Harrison, A., Barker, M. (2017). Investigating the diagnostic capabilities of monitored system parameters to support improvements in conveyor operation and maintenance. In First World Congress on Condition Monitoring. London, United Kingdom.
[16] Pang, Y. (2010). Intelligent belt conveyor monitoring and control (No. T2010/12). Netherlands TRAIL Research School.
[17] Massaro, A., Dipierro, G., Selicato, S., Cannella, E., Galiano, A., Saponaro, A. (2021). Intelligent quarry production monitoring risks and quality by artificial intelligence. In 2021 IEEE International Workshop on Metrology for Industry 4.0 & IoT (MetroInd4. 0&IoT), Rome, Italy, pp. 242-247. https://doi.org/10.1109/MetroInd4.0IoT51437.2021.9488469
[18] Olchówka, D., Rzeszowska, A., Jurdziak, L., Błażej, R. (2021). Statistical analysis and neural network in detecting steel cord failures in conveyor belts. Energies, 14(11): 3081. https://doi.org/10.3390/en14113081
[19] Stefaniak, P., Wodecki, J., Zimroz, R. (2018). Maintenance management of mining belt conveyor system based on data fusion and advanced analytics. In Advances in Technical Diagnostics: Proceedings of the 6th International Congress on Technical Diagnostic, ICDT2016, Gliwice, Poland, Springer International Publishing. Springer, Cham, pp. 465-476. https://doi.org/10.1007/978-3-319-62042-8_42
[20] Wang, M., Shen, K., Tai, C., Zhang, Q., Yang, Z., Guo, C. (2023). Research on fault diagnosis system for belt conveyor based on internet of things and the LightGBM model. Plos One, 18(3): e0277352. https://doi.org/10.1371/journal.pone.0277352
[21] Druzhinina, O.V., Masina, O.N., Petrov, A.A. (2021). Modeling of the belt conveyor control system using artificial intelligence methods. Journal of Physics: Conference Series, 2001(1): 012011. https://doi.org/10.1088/1742-6596/2001/1/012011
[22] Hao, N., Wang, X., Zhang, Y., Zhang, M., Sun, C., Hu, C. (2024). A machine vision-based edge detection method for belt lap of pipe belt conveyor. In Intelligent Equipment and Special Robots. IOS Press, pp. 80-86. https://doi.org/10.3233/ATDE240225
[23] Capelli, G., Verdi, D., Frigerio, I., Rashidian, N., Ficorilli, A., Grasso, V., Majidi, D., Gumbs, A.A., Spolverato, G. (2023). White paper: Ethics and trustworthiness of artificial intelligence in clinical surgery. Artificial Intelligence Surgery, 3(2): 111-122. http://doi.org/10.20517/ais.2023.04.
[24] Sun, X., Xiao, H., Meng, W. (2023). Research on the tension of steel cord conveyor belts based on transverse vibration modelling. Measurement Science and Technology, 35(3): 035028. http://doi.org/10.1088/1361-6501/ad180d
[25] Wu, L., Li, J., Zhang, H., Zhang, W., Huang, S. (2023). Research on a system for the diagnosis and localization of conveyor belt deviations in belt conveyors. Measurement Science and Technology, 35(3): 035110. https://doi.org/10.1088/1361-6501/ad0c48
[26] Hrabovský, L., Blata, J. (2024). The geometric shape of the transported material batches in the vertical branch of a belt conveyor. In MATEC Web of Conferences. EDP Sciences, 396: 07002. https://doi.org/10.1051/matecconf/202439607002
[27] Gavryukov, А., Kolesnikov, M., Zapryvoda, A., Lutsenko, V., Bondarchuk, O. (2024). Determining the mechanism for calculating the tension of a working conveyor belt during a change in the transportation length from medical mask waste. Eastern-European Journal of Enterprise Technologies, 128(7): 56-66. https://doi.org/10.15587/1729-4061.2024.300648.
[28] Belytskyi, Р. (2023). Influence оf load transportation speed by belt conveyor оn specific energy consumption іn conveyor transport. Journal оf Electrical and Power Engineering, 29: 38-42. https://doi.org/10.31474/2074-2630-2023-2-38-42
[29] Orlapu, A. (2024). The north-south conveyor belt model: Mapping the trajectory of international relations on the great ocean conveyor belt. Cogent Social Sciences, 10(1): 2297768
[30] Kirkire, M.S., Soman, A., Tikekar, A., Nevgi, V., Shinde, S., Bhise, A. (2022). Performance improvement in inclined belt conveyor for coal handling. In Biennial International Conference on Future Learning Aspects of Mechanical Engineering. Singapore: Springer Nature Singapore, pp. 497-510. https://doi.org/10.1007/978-981-99-1894-2_42
[31] Zeng, F., Zhou, J., Wu, Q. (2023). A conveyor belt characterization information extraction method based on YOLOv5 and the skeleton method. Measurement Science and Technology, 34(11): 115006. http://doi.org/10.1088/1361-6501/ace842
[32] Singh, U., Rahman, A., Mahato, S.K., Surya, S., Bajaj, M., Jurado, F., Kamel, S. (2022). Design of an automatic sorting machine on the conveyor belt system based on product temperature. In 2022 IEEE International Conference on Automation/XXV Congress of the Chilean Association of Automatic Control (ICA-ACCA), Curicó, Chile, pp. 1-6. https://doi.org/10.1109/ICA-ACCA56767.2022.10006123
[33] Miao, D., Wang, Y., Yang, L., Wei, S. (2023). Foreign object detection method of conveyor belt based on improved Nanodet. IEEE Access, 11: 23046-23052. https://doi.org/10.1109/ACCESS.2023.3253624
[34] Li, T., Zhang, H. (2024). Design of motor condition monitoring system for belt conveyor in coal mine. Highlights in Science, Engineering and Technology, 81: 689-692. https://doi.org/10.54097/z4j84280
[35] Zhou, P., Sun, L., Zhou, G., Ma, T., Wang, H., Bi, W., Ma, G., Wang, W., Yan, X. (2024). A new embedded condition monitoring node for the idler roller of belt conveyor. IEEE Sensors Journal, 24(7): 10335-10346. https://doi.org/10.1109/JSEN.2024.3363905
[36] Semrád, K., Draganová, K. (2023). Implementation of magnetic markers for the diagnostics of conveyor belt transportation systems. Sustainability, 15(11): 8705. https://doi.org/10.3390/su15118705
[37] Suthar, H.H., Butala, A.M., Patel, M.K. (2023). Cost and quantity comparative оf DLCS and PCS tower structures іn conveyor belt systems. Journal оf Offshore Structure and Technology, 10(2): 24-32.
[38] Sakharwade, S.G., Jha, A., Motghare, K.L., Modak, J.P. (2022). Formulation of mathematical model for transient belt stretch for full load starting of belt conveyor system. AIP Conference Proceedings, 2670(1). https://doi.org/10.1063/5.0115981
[39] Gładysiewicz, L., Król, R., Kisielewski, W., Kaszuba, D. (2017). Experimental determination of belt conveyors artificial friction coefficient. Acta Montanistica Slovaca, 22(2): 206-214.
[40] Zhu, S., Liu, C. (2023). Current status of belt conveyor fault monitoring. In Applied Mathematics, Modeling and Computer Simulation. IOS Press, pp. 637-642. https://doi.org/10.3233/ATDE231004.
[41] Soares, J.L.L., Costa, T.B., de Moura, L.S., Sousa, W.D.S., Mesquita, A.L.A., Mesquita, A.L.A., Braga, D.D.S. (2023). Decision tree for feature selection to diagnosis of belt conveyor idler. In 27th ABCM International Congress of Mechanical Engineering, Florianópolis, Brazil.
[42] Ilanković, N., Živanić, D. (2022). Fundamentals оf conveyor belts. In Contemporary Trends and Innovations in the Textile Industry, Belgrade, Serbia.
[43] Sun, X., Wang, Y., Meng, W. (2022). Evaluation system of curved conveyor belt deviation state based on the ARIMA-LSTM combined prediction model. Machines, 10(11): 1042. https://doi.org/10.3390/machines10111042
[44] Zhou, Q., Gong, H., Du, G., Zhang, Y., He, H. (2022). Distributed permanent magnet direct-drive belt conveyor system and its control strategy. Energies, 15(22): 8699. https://doi.org/10.3390/en15228699
[45] Shi, H., Shi, C., Li, Z. (2024). Application research on adaptive control method of variable frequency tensioning system for large belt conveyor. In Intelligent Equipment and Special Robots, IOS Press. https://doi.org/10.3233/ATDE240271
[46] Marasova, D., Andrejiova, M., Grincova, A. (2023). Experimental study of the influence of the interaction of a conveyor belt support system on belt damage using video analysis. Applied Sciences, 13(13): 7935. https://doi.org/10.3390/app13137935.
[47] Wang, Y., Du, Y., Miao, C., Miao, D., Zheng, Y., Yang, D. (2024). Longitudinal tear detection method for conveyor belt based on multi-mode fusion. Wireless Networks, 30(4): 2839-2854. https://doi.org/10.1007/s11276-024-03693-6
[48] Liu, J., Zhang, H., Xiao, D. (2023). Research on anti-clogging of ore conveyor belt with static image based on improved Fast-SCNN and U-Net. Scientific Reports, 13(1): 17880. https://doi.org/10.1038/s41598-023-45186-0
[49] Miao, D., Wang, Y., Li, S. (2022). Sound-based improved densenet conveyor belt longitudinal tear detection. IEEE Access, 10: 123801-123808. https://doi.org/10.1109/ACCESS.2022.3224430.
[50] Oleh, P., Maksym, S. (2022). Stress analysis in a conveyor belt for the maxwell model of elastic element. Bulletin of the National Technical University Series: Dynamics and Strength of Machines, 81-93. https://doi.org/10.20998/2078-9130.2022.1.265430
[51] Ilanković, N., Živanić, D., Katona, M. (2023). Analysis оf the conveyor belt tensile properties. In VI International Scientific Conference "Contemporary Trends and Innovations in the Textile Industry", Belgrade, Serbia, pp. 173-180.
[52] Jangam, D.D., Gurav, V.B., Pawar, N.N., Khatave, P.B., Sutar, U.S. (2023). Automation оf fruit sorting process using conveyor belt system. International Journal for Research іn Applied Science and Engineering Technology, 11(6): 2630-2635. https://doi.org/10.22214/ijraset.2023.53976
[53] Bałaga, D., Kalita, M., Siegmund, M., Nieśpiałowski, K., Bartoszek, S., Bortnowski, P., Ozdoba, M., Walentek, A., Gajdzik, B. (2023). Determining and verifying the operating parameters of suppression nozzles for belt conveyor drives. Energies, 16(16), 6077. https://doi.org/10.3390/en16166077
[54] Soares, J.L., Costa, T.B., Moura, L.S., Sousa, W.S., Mesquita, A.L., Mesquita, A.L., de Figueiredo, J.M.S., Braga, D.S. (2024). Fault diagnosis of belt conveyor idlers based on gradient boosting decision tree. The International Journal of Advanced Manufacturing Technology, 132: 3479-3488. https://doi.org/10.1007/s00170-024-13549-0
[55] Köken, E. (2022). Assessment of installed power for inclined belt conveyors using genetic algorithm and artificial neural networks. Konya Journal of Engineering Sciences, 10(2): 468-478. https://doi.org/10.36306/konjes.1085608
[56] Zhou, L., Wu, Z., Li, Y., Yao, H., Liu, Y., Yuan, Y., Meng, W., Han, L., Cao, X. (2023). Temperature characteristics of indentation rolling resistance of belt conveyor. Journal of Mechanical Science and Technology, 37(8), 4125-4135. https://doi.org/10.1007/s12206-023-0729-6
[57] Marijić, J., Vilić, M., Grgić, I., Karakašić, M., Ivandić, Ž. (2023). Product development of conveyor belts system with a focus on structural features. In 21st Natural Gas, Heat and Water Conference, 14th International Natural Gas, Heat and Water Conference, pp. 181-189.
[58] Lukanova, G., Ilieva, G. (2019). Robots, artificial intelligence, and service automation in hotels. In: Robots, Artificial Intelligence, and Service Automation in Travel, Tourism and Hospitality, Emerald Publishing Limited, pp. 157-183.
[59] Jiang, X., Namokel, M., Wu, J., Jiang, W. (2024). Control system design for belt conveyor. In Advanced Manufacturing and Automation XIII, Singapore, pp. 538-546.
[60] Yuan, C., Zhang, Y., Zhang, M., Wang, Z., Zhang, D. (2024). Belt rotation in the pipe conveyor: Research on a detection method based on image processing. In: Intelligent Equipment and Special Robots, IOS Press, pp. 93-99.
