Abstract
The quality of sliver at the draw frame stage is a decisive factor in achieving yarn uniformity, particularly in air-jet (Vortex, MVS), Open-end rotor, and friction spinning, where drawn slivers are passed directly into the spinning unit without a roving frame. Current auto-levelling systems enhance sliver evenness and productivity but can only be used for mass control, not to measure more detailed fiber-level properties. Therefore, the current auto-levelling technologies cannot be completely extended to high-level decision-making, fault diagnosis in real-time, and intelligent optimization of draw frame passages in accordance with the real sliver quality. Traditional methods of assessing slivers rely on offline testing with equipment such as the USTER Tester, which is highly precise but incapable of identifying defects during manufacturing. In this review, the author evaluates current technologies for monitoring fibers, slivers, and yarns, including optical, capacitive, and vision-based systems, as well as IoT-enabled platforms in modern spinning mills. An overall evaluation of the existing limitations shows a significant gap in real-time monitoring of sliver quality. To satisfy this requirement, a theoretical framework of an online sliver testing system is suggested to be incorporated at the draw frame delivery. The system also has integrated high-quality sensors, real-time sensor processing, machine sensor integration, and smart feedback control. The review illustrates the possibilities that such a system has to improve the draw frame operation, minimize scraps, enhance the uniformity of yarns, and facilitate Industry 4.0 changeover in spinning processes. This review not only summarizes existing sliver monitoring technologies but also defines the technological gap between current draw frame control and the requirements of real-time intelligent monitoring in spinning, proposing a feasible framework for future industrial implementation. The suggested solution not only improves the quality evaluation in real-time and process efficiency but also outlines the basis of AI-based manufacturing systems in the textile industry in the future.
Introduction
Sliver quality is very important in the spinning workflow, where the variability that is added at the draw frame stage directly affects the yarn form, fiber positioning, and the end product of the textile.1,2 The sliver quality is indicated by sliver irregularity, which is quantitatively expressed as the coefficient of variation (CV%) of sliver, which is the ratio of the standard deviation of the sliver to its mean linear density multiplied by 100.
CV% = (Standard Deviation/Mean) × 100. This parameter also provides a normalized measure of mass variation along the length of the sliver, which is widely used to evaluate sliver quality in the textile industry. These sliver defects can be systematically categorized into (i) mass variation (short- and long-term irregularity), (ii) periodic faults—small clusters of entangled fibers that occur in the sliver due to instability in the fiber control during drafting, (iii) neps—small clusters of entangled fibers, affecting the sliver appearance and performance, and (iv) drafting waves—due to instability in the fiber control during the drafting process. Though these measurable quality indicators exist, the use of industrial practices is predominantly on the basis of the offline testing, which leads to the late detection of defects. The delay causes faults to spread to the next process, causing loss of materials, lower yarn quality, and lower overall process efficiency. Hence, real-time and quantitative monitoring systems that are able to identify and categorize sliver irregularities during the manufacturing process are extremely important. Figure 1 shows the schematic of the spinning preparation process flow from carding through the draw frame to the spinning unit. Schematic illustration of the spinning preparation process flow, including carding, draw frame alignment, auto-leveling, and spinning stages. The figure is intended for conceptual visualization based on existing literature and does not represent original experimental data. Generated with AI-assisted tools and refined by the authors.
Air-jet spinning systems are popular in contemporary spinning plants because they are highly productive, labor-saving, and capable of producing relatively consistent-quality yarns.3–5 In the air-jet spinning technique, the drawn sliver is supplied directly to the draw frame and then to the air-jet machine without a roving frame. 6 Nevertheless, even with technological advancements in draw frame automation,7,8 sliver uniformity can still not be monitored during delivery (on-site). Traditionally, the quality of slivers is tested periodically using tools such as the USTER® Tester, which measures the mass of slivers, short fibers, neps, and periodic imperfections.9,10 Although this instrument is highly accurate, its application has a few disadvantages for spinning mills: they are expensive equipment, and they cannot be continuously monitored, and it must interrupt production or divert samples to offline analysis. Such limitations delay the detection of classified defects such as mass variation, periodic faults, neps, and drafting waves, allowing these irregularities to propagate into subsequent processes. As a result, faults that are added during the draw frame are often not detected until after the yarn testing, and this results in a high level of material wastage, inconsistency of the yarn characteristics, and a low level of efficiency during the process. Several studies have attempted in the last few years to examine sensor-based and machine-vision methods to measure the quality of textile materials in real-time, including the measurement of unevenness, hairiness, defects, and structural parameters of the yarn using optical, capacitive, and image-based methods.11–17 The principles behind optical sensing systems are based on light attenuation or scattering; changes in sliver density, diameter, or surface irregularities cause measurable changes in transmitted or reflected light intensity. In capacitive sensing systems, the change in dielectric constant of the sliver as it moves between the electrodes changes the capacitance signal, with the amount of the change being proportional to fiber density. These vision-based systems use high-resolution imaging and edge detection and image-processing techniques (such as Canny edge detection) to obtain geometrical and structural characteristics, including sliver width, contour variation, and defect distribution. Optical and capacitive systems have reasonable spatial resolution of sub-millimetre to millimetre scale, and good response times of millisecond order that are adequate for high-speed processing. Although vision systems can provide higher spatial resolution (pixel-level detection), they have a relatively long response time because of the acquisition and processing time for images. Although they all have these abilities, they have a common weakness, i.e., they can only extract external or bulk mass-related features and cannot directly solve internal fiber structure, such as fiber parallelization, hook configuration, and hidden fiber entanglements. This restricts their effectiveness in providing a complete structural quality assessment of sliver in real-time industrial environments.
In a complete spinning preparation line, sliver quality is not only dependent on the draw frame but also strongly influenced by upstream carding and downstream spinning stages.18–20 The opening of the fibres, the removal of neps, and the elimination of short fibres occur during carding, and the irregularities in the fibre structure that remain are emphasized during spinning.21–26 Hence, to ensure complete quality assurance in modern spinning mills, a holistic monitoring approach for the integration of the processes of carding, drawing, and spinning is crucial.27,28 Despite the fact that modern draw frames have auto-leveling and simple process monitoring systems, they are only capable of mass variation control and indirect process indicators. Although these systems can give partial information regarding drafting behavior, it lacks in-depth and real-time assessment of critical fiber level parameters, including fiber parallelization, hook configurations, nep forming, and fiber alignments (migrations) within drafting zones. Therefore, they are still not able to assist higher-level decision-making, including the optimization of the quantity of passages during the draw frame in accordance with the real quality of the sliver’s structure. Apart from Europe and Asia, cotton-heavy production areas in Africa and the Americas have been using the semi-automated draw frame and monitoring systems more, which have been introduced. Similar issues in the quality control of sliver during production are emphasized here in the industrial setups, particularly at high speeds and at an economical cost.
The study specifically related to monitoring sliver quality in real time at draw frame delivery, however, is scarce. A sliver testing unit directly incorporated into the draw frame delivery side is an exciting prospect for meeting present industrial requirements. This would allow the constant monitoring of sliver linear density, regular malfunctions, mass change, and drafting anomalies without stopping the machine, as the problem could be corrected immediately in the course of production.
This review aims to develop a synthesis of the current advances connected with draw frame control, technologies of sliver quality evaluation, and real-time monitoring plans, and defines the conceptual model of the newly developed online sliver testing system that is connected with the draw frame. This review intends to form a basis upon which future studies and materials applications in industry can be done by investigating the present constraints and technological describe individual sliver monitoring techniques in isolation. This review provides a comparative constraint and possible sensing and data-driven methodologies. The development of such a system of integration is capable of producing more reliable production, less waste, and a steady quality of sliver entering spinning systems.
Unlike previous studies that primarily focused on a critical evaluation of existing sensing, control, and digitalization approaches in draw frame systems. The analysis not only points to technological achievements but also to their current constraints, such as real-time structure monitoring, scalability to industrial speeds, and applicability to spinning systems that are based on Industry 4.0.
Significance of sliver quality in spinning preparation
Importance of sliver quality in spinning.
Sliver coefficient of variation (CV%) has a direct and proportional influence on yarn evenness (U%). Any increase in sliver CV% propagates through the drafting zone, amplifying yarn irregularity due to the conservation of mass variation during fiber attenuation. Consequently, higher sliver CV% results in increased yarn CV%, reduced uniformity, and higher imperfection index (IPI), particularly in high-speed spinning systems where drafting corrections are limited.
Traditional methods of sliver quality assessment
Conventionally, the quality of slivers is evaluated based on sliver offline testing periodically and then using a testing instrument like the USTER ® Tester.10,34 These systems are very precise in measuring the mass variation, neps, short-fiber content, and length distribution of the fibers. The USTER® Tester is based mainly on the capacitive measurement principle, which involves the sliver being passed between the capacitor plates, and variations in the dielectric properties affecting the capacitance signal, resulting from changes in fiber mass and density. This variation is directly related to sliver linear density and hence can be used to calculate the sliver irregularity, CV%, and other parameters. Although it is very accurate, the method has intrinsic measurable limitations. Testing is done on samples of material and, as a rule, at specified time intervals (e.g., every few hours or every batch of production); this means that real-time quality assessment is limited. Also, response time is not instantaneous, as data are analyzed and interpreted after collection, so it takes time to take corrective actions. This can result in product defects like periodic faults, neps, and drafting waves that may not be identified during production and carry over to subsequent processes, thereby creating waste and lowering yarn quality. As a result, capacitive testing is an effective method for statistical evaluation of performance offline, but is not capable of continuous monitoring and immediate correction of faults as required by today’s high-speed spinning systems.
Traditional sliver quality assessment methods.
Critical evaluation of offline sliver testing methods
Although offline systems such as the USTER® Tester provide high measurement accuracy, their role in modern high-speed spinning systems is fundamentally limited by their non-continuous nature. A critical analysis reveals that these systems are used for diagnosis rather than control measures. The result is that they are not able to record transitory defects like waves of drafting and short-term fluctuations in mass that may appear during actual production. Offline testing is not as timely as the new approaches to sensing online, and this results in a lag in taking corrective measures and defects being carried downstream. Offline testing is the industrial standard for quality checks and validation; however, it is not enough to allow a process to be optimised in real time.
Existing draw frame unit on spinning machine
In air-jet (vortex) spinning systems, the draw frame unit plays a particularly critical role because the drawn sliver is fed directly into the spinning machine without passing through an intermediate roving process. Hence, anything that is added or subtracted in weight at the draw frame stage is added or subtracted in the final yarn structure with little change. The high-speed draw frame with state-of-the-art auto leveling, accurate drafting control, and an electronic monitoring system is thus generally used on modern spinning lines to provide stable sliver quality at the delivery point. 35
Configuration and function of the draw frame unit
The draw frame unit used in spinning generally consists of multiple sliver feed cans, a drafting system (comprising break, main, and delivery rollers), an auto-leveling system, and a coiler mechanism that deposits the leveled sliver into cans positioned directly in front of the air-jet spinning machine.7,33,36,37 The primary objectives of this unit are fiber straightening, parallelization, blending through doubling, and reduction of mass variation in the sliver.
A quantitative expression of the drafting operation in the draw frame is the ratio of the delivery speed of the sliver to the feed speed, called the draft ratio. 38 Output speed (Draft = Delivery speed/Feed speed). The linear density uniformity of the sliver is critical in achieving good control of this ratio. The position in the drafting zone where the corrective action is applied to correct the mass variation is called the leveling action point (LAP) in an auto-leveling system. 39 However, because the point of measurement is remote from the point of control, a time lag is added to the control system. The response of the system has a phase lag between the input disturbance and the corrective output, which can be described using control system theory with a transfer function. At high delivery speeds (above 600 m/min), this response lag can become more important and can cause overcorrection or undercorrection of sliver mass variation. This means that there may be a lack of synchronization between sensing and actuation, which degrades leveling efficiency and may result in some residual unevenness in the delivered sliver. Accurate description of the dynamics of the systems, as well as the compensation of the transfer functions and delay of the system, is therefore very important for auto-level optimization in modern high-speed draw frames.40,41 Since air-jet spinning lacks a corrective roving stage, the draw frame becomes the final opportunity to stabilize sliver linear density and structural uniformity.
Critical limitation of the current draw frame architecture
Comparative analysis of modern draw frame systems shows that although they have improved the mechanical precision and accuracy of the auto-levelling system, their sensing architecture is still mass-based. This puts a structural restriction on measuring this because phenomena at the fiber level, such as the hook configuration, fiber migration, and local entanglement, cannot be measured directly. Furthermore, the distance from the sensing point to the point of correction adds a time delay, which is more and more important at high production rates. Indirect estimates of sliver quality, not actual structural measurement, are still required in even the most sophisticated systems that feature ANN-based leveling point optimization. This means that current draw frame systems are based upon statistical uniformity and not on physical fiber structure control.
Auto-leveling systems in modern draw frames
The majority of modern draw frames that are coupled to spinning machines use auto-leveling mechanisms that continuously control the change in mass of slivers by modifying drafting characteristics on command.
35
There are two approaches to control, which are open-loop and hybrid open-closed loop. Thickness of the slivers in open-loop systems is gauged between the drafting zone and the upstream of the drafting zone, and draft correction is applied at a predetermined Leveling Action Point (LAP). The hybrid systems are designed with an addition of open-loop mass correction along with feedback components to enhance responsiveness and stability, especially at high delivery speeds.
42
The LAP determination and optimization are also a significant factor that determines the efficiency of auto-leveling. Several studies have also shown that improper LAP setting may cause over-correction, time-delay errors, or a left-over unevenness in the supplied sliver. In a measure to solve this, there has been a successful use of artificial intelligence strategies, specifically artificial neural networks (ANNs), to forecast the best LAP locations that are based on the settings of the machine, fiber characteristics, and process dynamics. The same studies demonstrate that intelligent auto-leveling has a great effect in improving sliver CV% and minimizing faults caused by drafting, even when the delivery speed is more than 600 -800 m/min. At very high delivery speeds exceeding 800 m/min, the dynamic response of auto-leveling systems becomes increasingly limited due to sensor-actuator delay and phase lag in feedback loops. This results in reduced correction accuracy, increased overshoot, and residual unevenness in sliver mass distribution. Therefore, conventional control systems face fundamental performance constraints under ultra-high-speed industrial spinning conditions.8,40 Figure 2 presents the existing drafting system on the draw frame unit in the Vortex spinning system. Existing drafting system on the draw frame machine.
Intelligent control and digitalization of draw frames
In addition to traditional mechanical and electronic controls, intelligent control architectures are also emerging in modern draw frames. It has been proposed that ANN-based models, fuzzy logic controllers, and DSP-based control systems should be used to improve auto-leveling performance at high speeds and under high-draft conditions. 8 These systems enhance disturbance rejection, reduce the effects of dead time, and give a better sliver regulation compared to traditional proportional control systems. The latest trends within the industry corresponding to Industry 4.0 have also expanded the functionality of the draw frame units by providing digital monitoring and integration with the IoT. IoT-based draw frame monitoring architecture isolates functionality of real-time sensing and communication and cloud-based analytics with dual-processor architecture and lightweight protocols like MQTT.42–46 Remote supervision, continuous data logging, fault diagnosis, and combining with mill-wide manufacturing execution systems are all made possible through such systems. Nonetheless, these systems are concerned with process-control and monitoring signals rather than with high-resolution structural characterization of the sliver itself.
Limitations of existing draw frame units in spinning
Although significant advances have been made in the auto-leveling precision and digital control, current draw frame units employed in air-jet spinning are still reliant to a large extent on mass-variation sensing and contact-based measurement principles. They normally cannot detect finer structural data like sliver diameter profile, internal fiber arrangement, localized defects, or temporary drafting anomalies in real time at the point of delivery. In addition, auto-leveling systems are useful to correct overall changes in mass, but do not offer complete diagnostic data to support advanced analytics, predictive maintenance, or machine-learning-driven quality optimization. This means that whereas modern draw frames provide sliver uniformity that is acceptably good in air-jet spinning, there remains, nonetheless, a definite gap in technological advancement, which means that there is no non-contact, high-resolution, real-time sliver quality monitoring at the draw frame outlet. This is one of the reasons why the study proposes the inclusion of an online sliver monitoring unit (in addition to the current draw frame capability), which will not interfere with high-speed production. One of the biggest problems in connection with the real-time sliver monitoring system implementation is the retrofitting of the existing spinning mills. For integration, mechanical adaptation of draw frame units, installation of sensors without modification of the drafting geometry, and compatibility of the old control systems are necessary. While there are technical advantages, these limitations can prevent the industry from adopting the restrictions. Furthermore, although the modern draw frames have sensors on both the back and front drafting areas, which can provide information on process-based signals such as sliver thickness change, drafting force, or roller activities, they may only provide indirect and partial information about the fiber-level structural features. Current systems cannot accurately measure any of the parameters in real-time, such as fiber parallelization, fiber hook orientation, nep distribution, and finer fiber alignment. The existing draw frame technologies, therefore, cannot provide a consistent measure of the quality (index) of the sliver, which could be used consistently to make decisions at the process level, such as the need for a second or third draw frame passage. This limitation highlights the need for advanced and high-resolution monitoring systems to enable the seamless integration of real-time sliver structure and intelligent process optimization.
Technological development in online monitoring technology
Current textile studies have enabled faster realization of real-time monitoring techniques in various stages of spinning. Studies have explored: • Variation of yarn mass and defect sensors that are based on optical sensors.11,13–15,47 • Yarn mass variation sensors that are based on capacitive principles.13–15,47 • Image-processing applications to measure the yarn diameter, hairiness, and defects using Canny edge detection and clustering algorithms.12,16,17,48–53 • Control of material flow with the help of a closed-loop system for deep drawing.7,27,35,54,55 • Mainly, IoT and MQTT-based monitoring systems have been deployed to monitor spinning machines.42–46
These studies show the trend towards a growing transition to Industry 4.0-based solutions in the textile industry. Nevertheless, the transfer of these methods to sliver monitoring on the delivery side at the draw frame is not well examined.
A cross-technology comparison shows that optical systems have high spatial resolution and are sensitive to disturbances in the environment, like dust and vibration, which can reduce their industrial robustness. Stable and commonly used in mass production systems are capacitive systems, which are limited to bulk density measurement, but are not capable of resolving structural fiber orientation. In contrast to vision-based systems, these have a high computational cost and high latency, and are therefore not suitable for ultra-high-speed draw frame applications, where a good quality defect visualization is required. IoT-based architectures increase access to data and the connectivity of processes, but do not improve the sensing ability. Thus, the combination of the non-contact operation, real-time response, structural sensitivity, and industrial robustness cannot be met by any of the available technologies.
The imaging-based sliver inspection device (Patent No. WO1999034044A1) is one of the most technically relevant and first sliver monitoring devices that are patented.
56
It was an invention of a non-contact imaging system that could detect impurities in a moving sliver in real time and did not stop production. The patented design consists of: • Curved aluminum direction plates or transparent rollers that press the sliver softly to cleave the bundle of fibers, minimize air spaces, and enhance optical clarity. • A xenon stroboscopic light source, connected to a camera with a pulse generator to synchronize the light source with a CCD camera, allows frozen frame imaging even with very high sliver speeds. • Clear optical windows through which the transmitted or reflected light can pass through the compressed sliver. • High-resolution pixel array that produces digital images to analyze them. Such an image-processing unit is used to determine neps, seed coat fragments, trash particles, and fiber entanglements by finding clusters of dark pixels above a set threshold, shape, fuzziness, and density, and comparing the output to pre-programmed reference images.
The earlier patents tended to be targeted at impurity detection; more recent patents have examined using non-contact optical and sensor fusion to measure structural irregularity and estimate CV% of sliver. The systems, however, are still not widely used in industrial applications because of their complexity and speed limitations in high-spinning-throughput environments. The main innovation in the system is optical imaging as opposed to capacitive or mechanical sensitivity, which allows a deeper understanding of the sliver structure. It identifies internal defects- unlike the previous systems that identified surface contamination only. The machine could be fixed between the card and draw frame, or could be fixed on the existing lines without significant alteration, so that impurities could be detected online during the processing. The patent, however, puts more emphasis on impurity recognition rather than complete sliver-quality parameters (e.g., mass CV%, drafting faults, structural irregularity). This system also depends on compressing the sliver, so it is not very suitable in high-speed delivery of the draw-frame, where draft change must never be disrupted.
A significant advancement was made by Zhu et al. (2007), 8 who developed the high-speed auto-levelling system based on the combination of open-loop and closed-loop control and the use of fuzzy logic. Their hybrid solution tackled significant challenges in traditional systems, that is, dead-zone errors, sensitivity to disturbance, and poor performance at high drafting rates. They have made significant improvements in sliver CV% at velocities up to 800 m/min by incorporating fine servomotors and rapid analog control equipment. Although these advancements have been made, the conventional systems are still limited to mass-variation feedback, rely on physical-contact sensors, and cannot present high-resolution structural data or digital data that can be used in advanced analytics.
Liu et al. (2025) 35 also designed and experimentally tested a high-speed drawing frame-specific open-loop, fixed-length auto-leveling control system (ACS). They are based on a DSP-based architecture that incorporates the major modules, which include a touchscreen interface, sliver thickness sensor, speed sensor, servo drive, and an analog-to-digital converter. In this platform, the authors were able to apply auto-leveling algorithms, communication protocols, and real-time monitoring of sliver quality. The experimental validation showed that the proposed ACS has a significant effect on enhancing the quality of slivers. The CVm of the sliver had decreased by 93.10% compared to the unleveled sliver, which indicates a significant effect on the improvement of thickness uniformity. The system also reduced the root mean square error (RMSE) of sliver weight by 11.75% and the average coefficient of variation (CV%) by 5.59 when compared to an already existing advanced commercial ACS. These findings verify that the proposed open-loop fixed-length control method can offer high stability in sliver thickness detection and gives competitive results compared with sophisticated systems. Mostly, the study provides effective technical data towards the formulation of steady, high-performance domestic ACS solutions to the contemporary high-speed drawing frames.
The recent trends of Industry 4.0 have changed the emphasis to IoT-based draw frame monitoring. Research using MQTT communication, two-processor communication, and cloud-based data management reveals that remote supervision, fault diagnosis, and continuous data logging are possible. Cui et al. (2018) designed a high-speed drawing frame monitoring system on the Internet of Things (IoT) built on the Message Queue Telemetry Transport (MQTT) protocol,
42
which comprises drawing frame IoT terminal equipment, the cloud server, and PC or mobile terminal equipment. The main developments of IoT-Based Draw Frame Monitoring include: • Implementation of a standard online detection system that will not be related to the auto-leveler control system. • With a cortex-A7 and a Cortex-M4 of a sequence of i.MX7 dual-processor architecture, isolating real-time sensing functions and communication/cloud uploading. • Linux and FreeRTOS are used as the operating systems by the two processors, respectively, and the information transfer between the two is taken in the rpmsg pattern. • MQTT lightweight communication application for real-time data streaming to PCs and mobile devices. • Allowing remote monitoring, fault diagnosis, and data storage in a cloud-based analysis model. • Avoidance of wired industrial networks by adopting wireless IoT communication.
Existing sliver monitoring & draw frame control.
Comparison of draw frame control strategies.
Schematic representation of key sliver quality parameters, their common measurement techniques, and their general impact on yarn properties, as reported in the literature. The figure is intended for illustrative purposes to summarize existing knowledge and does not represent original experimental data or novel analytical results. Generated with AI-assisted tools and refined by the authors.
While the benefits of monitoring with the help of IoT are significant, the issues of cybersecurity and data scalability still remain serious challenges. Industrial systems need to implement protocols, encryption, and authentication methods to stop unauthorized access. Moreover, scalable cloud architectures and edge computing approaches are required to cope with the high number of sensors producing data in large-scale spinning mills.43,45
In comparison to commercial auto-leveling control systems (ACS), most research-based systems show a better CV% reduction and are still not integrated with structural monitoring and predictive analytics. However, there is a disconnect between what is achieved in the laboratory and the industrial results, with commercial systems optimized for industrial stability, while proposed academic systems are aimed at improving sensing and data-driven control.
To further clarify the strengths and limitations of existing approaches, a comparative evaluation of sliver monitoring technologies is necessary. Existing systems differ significantly in sensing principle, response time, and measurable parameters. Optical systems provide surface-level defect detection, capacitive systems measure bulk mass variation, while IoT-based systems focus on connectivity and process monitoring. However, none of these approaches simultaneously provide high-resolution structural, real-time, and non-contact measurement at the draw frame delivery point. In addressing the shortcomings of compression-based optical systems, laser scanning without contact and hyperspectral imaging might be proposed. Laser-based systems allow for accurate measurement of diameter and displacement without contact, and hyperspectral imaging allows for the spectral signatures of sliver fiber composition and variation, thereby providing greater insight of sliver quality at high-speed operation.
Research gap
The primary research gap can be defined as the absence of a real-time, high-resolution measurement system capable of simultaneously capturing both mass-related and structural parameters of sliver during the draw frame process. The current technologies mainly focus on measuring bulk variations in mass and on the detection of defects at the surface level, and do not offer any detailed information about the structure of the fibers. In particular, current systems do not provide any means to measure key parameters like fiber orientation, degree of parallelization, hook configuration, and local defects in the sliver cross-section. Furthermore, traditional approaches lack the spatial resolution and principles used to detect small-scale irregularities and transience in the high-speed production environment. Because of this, no integrated solution enables continuous high-accuracy monitoring of sliver mass uniformity and internal fiber arrangement, which is a major technological gap in the current spinning preparation process. On the whole, the literature demonstrates significant advances in the accuracy of control and digital monitoring; however, an obvious gap exists: there is no high-resolution, non-contact, real-time monitoring system of sliver quality that would be integrated into the draw frame outlet. The need underlines the basis of the proposed sliver testing unit development.
The existing constraints are: • Failure to have online sensors that are specifically developed to deal with thick, low-twist slivers • Inability to check the mass variation and defects at fast delivery rates • Lack of integration of machines and sensors with real-time control logic • Data-driven predictive maintenance of draw frames has been adopted poorly.
This gap provides the basis of the necessity of an integrated sliver testing unit that can be utilized at full production rates without machine stoppage.
Conceptual framework for an online sliver testing system
The online sliver testing unit, as proposed, will be mounted in the circumference of the delivery side of the draw frame, where quality assessment can be done instantly before the sliver goes through spinning. Figures 4 and 5 present the conceptual drafting-unit design and the conceptual framework of the draw frame unit, respectively. Conceptual drafting-unit design of the draw frame unit. Conceptual framework for the online sliver testing system.
System objectives
• Constant observation of the changes in sliver mass. • Faults in drafting, intermittent mistakes, and sudden quality variations are detected. • Integration with machine control to give alerts or automatic corrective action. • Real-time operator data storage and data visualization. • Non-contact or minimum contact measurement to prevent sliver damage.
Key functional components
The proposed online sliver monitoring system is implemented by a sensor fusion method, which combines optical sensing and capacitive sensing principles to provide holistic quality assessment. Within this framework, optical sensors measure the diameter variation and irregularity of surfaces by determining light attenuation or scattering, while capacitive sensors measure mass variation, with the capacitive changes caused by the dielectric changes. These complementary signals combine to provide more reliable measurement and the ability to assess sliver quality with multiple parameters. Outputs from the system include a coefficient of variation (CV%) of the mass and diameter, real-time diameter fluctuation profiles, and defect frequency (thick places, thin places, neps per unit length). Data in the system is sent in a predefined sequence: The raw signals are collected through the sensors, processed by digital signal processing and pattern recognition algorithms to obtain useful information, and sent to the control unit for on-line decision making. From the processed data, corrective actions like draft adjustment, machine slowdown or fault alerting can be automatically undertaken.
The integrated sensing–processing–control system is used to continuously monitor and intelligently optimize the draw frame operation in real time.
a. Sliver sensing module
The optical and capacitive sensing methods are available and being used, but new methods in the process of development, such as acoustic emission sensing and ultrasonic wave propagation analysis, are also considered for the deeper characterization of the fibers. Acoustic sensors can detect vibration signatures related to drafting instability, and ultrasonic techniques can give information on changes in internal fiber compactness and structural density. The combination of multiple modalities (optical, capacitive, acoustic/ultrasonic) in the sensing system could also improve the detection accuracy.
The possible types of sensors would be: • Capacitive sensors (mass variation detection), • Optical displacement/laser sensors (diameter fluctuation), • 2D/3D camera vision systems (sliver profile mapping), • Tension sensors (find drafting waves or slippage) using load cells.
b. Signal processing/feature extraction
The signal processing and data analysis stage is crucial to the ability to convert raw sensor measurements into meaningful quality measures. The first approach to noise reduction is to apply filters, for example, low-pass filters, to reduce the influence of noise components at high frequencies, while maintaining the signals of the sliver variations. Afterward, feature extraction techniques are used to detect important irregularities within the signal. For example, thick and thin places are quantified using peak detection algorithms for finding local maxima and minima above a set threshold. Other statistical parameters like mean, standard deviation, and coefficient of variation (CV%) are also calculated for continuous quality evaluation. Moreover, machine learning techniques can be adopted to improve defect identification and classification. Defects can be divided into many categories, such as neps, periodic faults, and drafting waves, and the classification can be realized by training supervised classification models using labeled datasets, including support vector machines (SVM), artificial neural networks (ANN), and decision trees. This signal filtering, feature extraction, and intelligent classification make it possible to do a real-time quality analysis of slivers in modern monitoring systems without any error. In the case of predictive maintenance, machine learning models can be used for temporal sliver quality data analysis, such as Long Short-Term Memory (LSTM) networks and Random Forest classifiers. The models can learn and predict possible drafting errors, roller wear, and irregularity of sliver by taking advantage of historical sensor patterns. This will help to detect and predict the fault in a smart spinning system at an early stage and undertake maintenance based on the condition of the system. The parameters that will be analyzed in the system will be: • CV% of mass or diameter, • Thick and thin places, • Formation of neps.
Machine learning models can be used to identify abnormal trends or anticipate future failures.
c. Feedback and real-time analysis
The feedback and control system is based on the closed-loop control principle, with the parameters of sliver quality measured and compared continuously with the target value. The input to the control system is the difference between the measured value and the reference set point, called an error signal. From this error, the control actions are calculated, and process stability is maintained through minimising deviations. Common control measures are adjustments to the draft by changing the speed of the front or back rollers, a change in roller pressure, and dynamic control of the speed of the draw frame delivery system. The control system response time and stability characteristics are important factors in determining its effectiveness. The system should respond quickly enough to correct disturbances promptly, but very rapid corrections can cause oscillations or instability to occur in the system. Therefore, proper tuning of control parameters is required to ensure stable operation, minimize overshoot, and achieve steady-state accuracy. This closed-loop mechanism enables continuous correction of sliver irregularities and enhances overall process consistency in high-speed spinning operations.
Processed information will be represented by: • Local HMI touchscreen, • SCADA dashboard, • IoT/MQTT remote-monitoring cloud platform.
d. Draw frame control
The Draw Frame Control can be used with Integration. The system may enable: • Fault automatic machine slowdown due to faults, • Roller pressure/tension adjustments, • Preventive maintenance notifications.
Advantages and industrial impact
The advantages of the proposed online sliver testing system to the industry include: • Eliminates offline testing, which assures the instantaneous identification of quality deviations. • Reduces waste through detection of problems prior to sliver being sent to spinning. • Enhances consistency of yarns, particularly in air-jet spinning, which does not have a roving step. • Improves machine performance since there is no need to stop to sample sometimes. • Lessens the reliance on labor through the use of automated and round-the-clock monitoring. • Enables Industry 4.0 transformation through sensor and data analytics integration and cloud connectivity. This system can significantly enhance the reliability and stability of the high-speed spinning process. Table 5 represents the proposed conceptual online sliver monitoring system. Proposed conceptual online sliver monitoring system.
Another significant implication of the proposed online sliver monitoring system is that it has the potential to facilitate intelligent decision-making on the number of draw frame passages to use during spinning preparation. The system can facilitate a more detailed assessment of sliver quality than traditional mass-based indicators by enabling real-time analysis of important structural parameters, such as fiber parallelization, fiber hooks, neps, and total fiber alignment, especially when monitoring across various drafting zones. Depending on such high-resolution quality feedback, the system may help operators or automated control logic decide whether the sliver has already reached the desired quality level following the initial passage through the draw frame, or whether it requires further passages (e.g., second or third draw frame). In case the monitored parameters show an adequate level of fiber straightening, low hook content, and low nep content, the following draw frame processes can be minimized or omitted, resulting in energy, time, and cost of processing savings. On the other hand, poor alignment of fibers or increased defect rates would be the reason to continue with the drawing process in order to guarantee the final quality of the yarn. Therefore, the suggested system can not only increase the quality monitoring in real time, but also present a data-driven structure for the optimization of process routes in spinning mills, which cannot be done with the current conventional draw frame control systems. Figure 6 shows the comparison of the traditional and online sliver monitoring system. Comparison of traditional and online sliver monitoring.
Human–machine interaction is also important to the successful implementation of real-time monitoring systems. Operators need to be properly trained to understand the outputs of predictive analytics, digital dashboards, and alarms. To properly implement automation and the use of AI-based recommendations in industrial settings, it is vital that these are accepted and that trust in the recommendations is developed. From an economic perspective, the use of RT sliver monitoring systems could require more initial investment for sensors and IoT integration; however, the benefits of implementing this system are that material wastage will be reduced, rejection rates will be lowered, downtime will be reduced, and yarn consistency will improve. All these factors work together to enhance productivity and profitability (ROI) in high-speed spinning mills.
In terms of sustainability, sliver performance monitoring reduces raw material waste, reduces energy consumption and related reprocessing cycles, and reduces the production of defective sliver. These help to achieve the Sustainable manufacturing goals by reducing the impact of spinning activities on the environment.44,46
Technological advancement compared with existing draw frame control
Existing draw frame control systems in spinning primarily rely on auto-levelling mechanisms based on mass-variation sensing and contact-type measurement principles. These systems effectively reduce average sliver irregularity through open-loop or hybrid control strategies; however, they are limited in their ability to provide high-resolution structural information, real-time defect localization, and predictive quality diagnostics.
The technological framework proposed in this study differs from conventional control units in several key aspects: (i) Non-contact multi-sensor monitoring integrating optical, capacitive, and vision-based measurements instead of single-parameter mass sensing; (ii) Real-time data processing and anomaly detection using digital signal processing and machine-learning-assisted analysis rather than only feedback draft correction; (iii) Integration with IoT-enabled supervisory platforms allowing remote diagnostics, historical trend analysis, and predictive maintenance; and (iv) Placement of the monitoring unit directly at the draw frame delivery, enabling quality verification immediately before spinning, which is not achievable with current auto-levelling systems.
Thus, the proposed system is an application of the new sliver quality monitoring approach based on data assimilation and control theory, which moves beyond the conventional mass-regulation-based approach of current draw frame technologies.
Industrial necessity for real-time sliver monitoring in spinning
The performance of the proposed monitoring system has a significant effect from an industrial feasibility point of view, as sensor response characteristics at high production speeds play an important role. For delivery speeds greater than 600 m/min, the sensors have to possess short response times (milliseconds) and high sampling rates (usually kHz) in order to properly detect sliver fluctuations and localized defects during the transitory phases. If sampling is too slow, there can be aliasing effects, and important defect information can be lost. Furthermore, calibration of the sensing units is crucial to guarantee accurate measurements and consistency for different sliver types and conditions. These calibration techniques can be verified by testing reference materials, normalizing baseline signals, and recalibrating periodically under controlled conditions. Machine vibrations, airborne dust, changes in temperature, and humidity can all have a significant impact on the performance of a sensor, introducing noise and signal distortion. In real industrial environments, therefore, dependable operation requires robust system design, which includes the use of vibration isolation, protective enclosures, and adaptive signal compensation techniques. These considerations are critical for ensuring the practical applicability and long-term stability of the proposed monitoring system in high-speed spinning mills.35,42
The industrial relevance of developing an online sliver monitoring system is strongly linked to the operational characteristics of spinning, where the drawn sliver is directly supplied to the spinning unit without an intermediate roving stage. Consequently, any undetected irregularity at the draw frame delivery propagates directly into the yarn structure, leading to yarn quality variability, increased waste, and production inefficiency.
Although modern auto-levelling draw frames improve average mass uniformity, they do not provide continuous structural quality verification or early fault diagnosis. Industrial spinning mills increasingly demand real-time quality assurance, reduced material loss, predictive maintenance, and Industry 4.0-compatible monitoring. These requirements cannot be fully satisfied by existing draw frame control architectures.
Hence, the development of a non-contact, real-time, intelligent sliver monitoring system is not only technologically meaningful but also industrially necessary for high-speed spinning environments.
Scope of the present review
The present study is a conceptual and technological review aiming to synthesize existing knowledge and propose a future framework for online sliver monitoring. Experimental implementation is beyond the scope of this article and is recommended as a direction for future research. Some of the key performance indicators (KPIs) used when assessing sliver monitoring systems are the decreases in sliver CV%, defect rate (thick places, thin places, neps), machine downtime because of fault detection and correction efficiency, and improvement in yarn uniformity (U%). These are metrics that will give standardisation to the performance of a system. It should be pointed out that the proposed framework is still a concept and has not undergone full-scale industrial pilot testing. In the future, the implementation in operational spinning mills should be aimed at assessing the practical performance, robustness under production conditions, and feasibility of integration with the existing draw frame in operational conditions.
Conclusion
With the growing use of high-speed spinning systems, there has been a growing pressure on maintaining a steady and consistent quality of sliver products. Conventional offline sliver testing techniques cannot provide real-time information or immediate corrective action, leading to production inefficiencies and quality variability. This review also shows a dire need for an online sliver-testing system that can be used for continuous measurement on the draw-frame delivery side. Comprising the contributions of breakthroughs in sensor technologies, machine vision, signal processing, and IoT-driven monitoring platforms, this review creates a technological background for creating intelligent sliver monitoring systems. The given conceptual framework provides a good direction for embedding non-contact sensing, real-time data analysis, and automated feedback control into the current draw frame functions. These advancements can greatly decrease waste, enhance uniformity of the yarns, and reinforce the reliability of the processes in modern spinning mills. This review identifies the important need for existing draw frame systems to measure sliver quality in real time, which currently does not have structural, non-contact, and high-resolution measurement capability. It highlights the need for further research and development that would involve the incorporation of multi-sensor, data-driven monitoring systems in order to allow for real-time quality assessment and process control in spinning preparation.
Future development should be directed to prototype building, sensor calibration for thick slivers, as well as interconnection with industrial draw frame systems to check their functionality in real production. Furthermore, the suggested system is compatible with the future vision of AI-enabled textile manufacturing, such as real-time data collection, intelligent analysis, and adaptive process control. Machine learning algorithms combined with real-time sensor information can enable predictive quality control, fault detection and diagnosis, and self-decision making in spinning processes. It is also aimed at adaptive control systems based on artificial intelligence, modelling of draw frame processes via digital twins, and advanced multi-sensor fusion architectures. Real-time sensing combined with virtual process simulation can pave the way for predictive optimization, autonomous decision-making, and self-correcting spinning systems in the future of textile production.
Although the steps toward automation and digital monitoring have come a long way, the existing draw frame systems still suffer from the use of indirect measurement principles and a lag time in the feedback. Therefore, an improvement in the uniformity of the sliver does not always reflect an improvement in the actual fiber-level structure of the sliver, demonstrating a lack of correlation between process control and fiber material science knowledge.
Footnotes
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The authors confirm that the data supporting the findings of this study are available within the article.
AI-assisted content disclosure
In this study, artificial intelligence (AI) tools (Google Gemini) were used solely for the generation of representative schematic illustrations (Figures 1 and 
) to support the visualization of concepts discussed in the literature. These figures are intended for illustrative purposes only and do not represent original experimental data, analytical results, or novel research findings.
All AI-assisted outputs were carefully reviewed, modified, and validated by the authors to ensure accuracy and consistency with the cited sources. No AI tools were used in the development of the research methodology, data analysis, interpretation of results, or formulation of conclusions.