Manufacturing Line Research Articles

A Quantified Analysis of Lean First Then Automate for a Synergetic Effect

ABSTRACTLean manufacturing has become a buzzword for the past few years due to its capability to reduce waste and make the system more efficient. To meet the numerous demands of consumers, companies are moving toward Industry 4.0, which can be accomplished through automation. Prior research demonstrated a conceptual integration of lean tools' effects into automation, but it lacked a quantitative analysis. The purpose of this study is to ascertain the effects of implementing lean tools prior to automation. To get real‐time data for the analysis of applying lean and automation, an empirical study was conducted on a manufacturing line. Using ARENA simulation software, four distinct simulation models were created using the data: the current assembly line model, the model with a lean tool (Inline Quality Check), the model with only automation, and the model with lean applied first and automation applied later. The productivity of the assembly line is used to compare the results. The comparative analysis demonstrated that applying lean tools prior to automation increases productivity in automated lines. This study could assist businesses understand that, before implementing automation, they should use lean technologies to improve system efficiency. An effective system can become considerably more efficient by being automated.

Effect of manufacturing tolerances on Micro-CPV assemblies: A quantitative approach based on statistical modeling

In micro-concentrator photovoltaics (micro-CPV) minimized components as cells (<1 × 1 mm2) and lenses are used, promising significant cost reductions through parallel manufacturing and reduced material volumes. However, tolerances, such as deviations from nominal size, geometry or position, impact module performance, especially for non-ideal alignment towards the sun. To study the interplay of different, independent tolerances and their effects on current generation, a comprehensive parameter study is practically not feasible, because of the vast number of possible combinations. In this work, we introduce a novel method for assessing tolerances by employing a Monte-Carlo approach to randomly select and combine tolerances in a cell-lens unit. It allows to identify relevant tolerances and quantitatively assess their influence on module performance, namely optical efficiency, and photocurrent as function of angle of incidence and, thus, acceptance angle. We apply the model to a micro-CPV module developed at Fraunhofer ISE and use tolerance distributions based on measurements. We find that the most crucial parameter is the position of the secondary optical element. Given the measured tolerance distributions, the acceptance angles for 90 % of the cases are above 0.5° for 10 % current loss. The developed approach is a crucial tool for identifying and assessing critical tolerances within a manufacturing line, facilitating techno-economic optimization of design and manufacturing processes.

A review on protein utilization and its interactions with carbohydrate and lipid from a molecular perspective in aquaculture: An implication beyond growth.

The world's increasing need for protein faces challenges in aquaculture production. New applications and tools will need to be added at every stage of the manufacturing line to attain this expansion sustainably, safely, and effectively. Utilizing experimental methods to increase aquatic animal production has become more common as aquatic biotechnology has advanced. High-throughput omics technologies have been introduced to address these issues, including transcriptomic, metabolomic, proteomic, and genomes. But it also faces many difficulties, like other food manufacturing industries. One of the best and most durable approaches to address these issues is probably to understand nutritional requirements and modify diet based on need. Molecular approaches are a subset of multiomics technology. Previously, most of the published work was devoted to the biochemical aspects of protein-lipid interactions in biological systems. In this review, we explore this idea and highlight various works that fall under the umbrella of nutrigenomics, with a particular emphasis on protein utilization and its interactions with carbohydrates and lipids.

Establishment of a semi-continuous nano-production line using the Microfluidizer® technology for the fabrication of lipid-based nanoparticles part 1: Screening of critical parameters and design of experiment optimization studies

A variety of strategies for producing high-quality nanoparticles have been reported in recent years. Batch-based bottom-up and top-down technologies are generally the most efficient methods, but present a number of challenges, particularly in terms of variability, safety, sustainability and large-scale production. In this study, a scalable, semi-continuous production line was built by connecting individual processing units, including a high shear mixing device, the Microfluidizer® technology and a cooling system. Each unit was equipped with an adequate temperature control to allow solvent-free production of solid lipid nanoparticles (consisting of Precirol® ATO 5 or Gelucire® 43/01) and nanostructured lipid carriers (additionally comprising Labrafac™ lipophile WL 1349). Subsequently, critical formulation parameters and critical process parameters (CPPs) of the individual processing units and their effects on particle size (i.e., critical quality attribute (CQA)) were investigated to identify appropriate input parameters for the subsequent Design of Experiment (DoE) studies conducted after linking the process units to a semi-continuous production line. For particle size monitoring, spatially resolved dynamic light scattering (SR-DLS) measurements were conducted and compared to standard DLS measurements to evaluate the applicability of SR-DLS as an inline monitoring tool. It was found that matrix composition, emulsifier concentration, pressure and number of cycles when processing through Microfluidizer® processor were the most influencing parameters. By optimizing these parameters, five-times higher throughputs could be achieved by the semi-continuous manufacturing line. In addition, the particle size measurements with SR-DLS confirmed the feasibility of implementing this technology for real-time particle size monitoring as an important safety factor in quality control.

Characterisation and manufacturing methods of material extrusion 3D printing composite filaments based on polylactide and nanohydroxyapatite

As the pace of scientific research and technological advancement accelerates, there is an increasing demand for rapid on-site prototyping with specific materials. Additive manufacturing through material extrusion 3D printing has emerged as an economically viable solution to this need. The production of custom filaments for this process frequently depends on low-volume filament manufacturing lines. A significant challenge in such production processes is maintaining the quality of the filament throughout its entire length, as this has a considerable impact on the quality of the final printed object. To ensure consistent quality, it is essential to monitor and manage a range of essential properties, including diameter consistency, material homogeneity and compound ratio. This study presents the implementation of a real-time filament quality monitor (R-FQM), which provides detailed insight into the longitudinal characteristics of the filament during the manufacturing process and its crystallinity analysis. In conjunction with a new proof-of-concept drawing process for reducing filament diameter by multi stage drawing die solid-state extrusion, this approach addresses the challenge of maintaining filaments quality. Several PLA-based filaments were produced for research purposes: pure PLA and composites containing 5 and 10wt% nanohydroxyapatite. Produced composite filaments had significantly greater diameter variation compared to pure PLA (largest for 5% filler ± 0.29mm). The diameter of the resulting filaments was reduced in the range of 2.5-1.4mm using the SSE approach. In addition to the R-FQM method, the filaments were characterised by structural, thermal and mechanical property analysis. The analyses showed that the 10% hydroxyapatite content in the filament subjected to the SSE process can lead to significant weakening of the material. However, all the filaments were suitable for 3D printing for biomedical applications. The research carried out demonstrated the complementarity of the R-FQM method with classical methods, confirming the effectiveness of the approach described.

A cradle-to-gate life cycle assessment for clean hydrogen gas production pathway using the CeO2/Ce2O3-based redox thermochemical cycle

The present study is conducted to develop a two-steps redox thermochemical cycle based on combined ceria reduction and methane reforming through the CeO2/Ce2O3 redox pair. This cycle incorporates simultaneous water splitting for hydrogen production. To enhance the accuracy and reliability of the inventory data, the system is simulated using the Aspen Plus software package. Also, the goal of this work is to evaluate the carbon footprint of thermochemical based hydrogen production under different scenarios and to compare it with traditional technologies like steam methane reforming (SMR). The analysis takes into account both the plant manufacturing line (reactors), fuel processing (H2 production), redox particles oxygen carrier manufacturing (OC), H2 compression (power consumption) and H2 storage (pipeline manufacturing). The openLCA software with the European database environmental footprint is employed to compute the total carbon emissions of this thermochemical cycle during its overall life cycle. The study concludes that the base case scenario of this system has an overall GWP of 1751.14 g CO2 eq./kg H2 respectively. In the overall GWP, the share of plant construction and OC production was found to be around 1.67%, and 15.78% respectively. The conventional steam methane reforming is loaded with an overall GWP of 4472.83 g CO2 eq./kg H2. In terms of renewable energy input, the base case scenario of this hydrogen production system results in 1465.73 g CO2 eq./kg H2 with thermal energy produced from renewable sources. Comparative life cycle assessment (LCA) demonstrates that base case of CeO2/Ce2O3 has an 85.41% lower GWP emphasizing its environmental advantages over the SMR and Fe3O4/FeO redox pair carries an 84.06% lower GWP as compared to the conventional SMR with 1483.55 g CO2 eq./kg H2. The present study concludes that the combined ceria reduction and methane reforming chemical looping process with simultaneous water splitting may be a more promising option to produce hydrogen using renewable energy.

Improving Continuous Loss-in-weight Feeding Accuracy by a Novel Hopper Design

PurposePowder feeding is a vital unit operation in the continuous manufacturing of pharmaceutical products. Loss-in-weight twin-screw feeders are commonly used in continuous manufacturing lines. The feeding performance, i.e., the accuracy and consistency of feeding, influences the content uniformity of the final drug product. In this study, a redesigned hopper for K-Tron KT20 twin-screw feeder was designed, implemented, and investigated to improve feeding performance. The basic idea was to design the hopper in such a way that the stresses on top of the screw entrance are independent of the fill level.MethodsOur study compared the novel system to that of the original cylindrical hopper. The effect of the redesigned hopper on the start-up, the process’ sensitivity to refill level, and the refill portion size were studied. A free-flowing, barely compressible powder and a compressible blend were used in this study to evaluate the feeding performance and the refill effects.ResultsWhen using compressible powders, the results showed a larger process window for refilling in the redesigned hopper and a lower refill-level sensitivity compared to the original hopper.ConclusionThe homogenization of stresses in the redesigned hopper allowed operating at lower refill levels, and variability, especially due to refilling, was reduced. This study demonstrates novel design approaches for successfully implementing hard-to-feed materials in continuous manufacturing.

The effect of process parameters on a continuous blending process monitored in-line by near-infrared spectroscopy

The continuous feeding-mixing system ensures the composition uniformity down to the tableting continuous manufacturing line so that a quality end-product is consistently delivered. Near-infrared spectroscopy (NIRS) enables in-line assessment of the blend's critical quality attributes in real-time. In this study, the effect of total feed rate and impeller speed on the continuous blending process monitored in-line by NIRS was examined by principal component analysis (PCA), ANOVA simultaneous component analysis (ASCA) and partial least squares (PLS) regression. Process data were generated by a factorial experimental design with process parameters and a constant formulation comprised of: 30 % (wt/wt) ibuprofen, 67.5 % (wt/wt) microcrystalline cellulose, 2 % (wt/wt) of sodium starch glycolate and 0.5 % (wt/wt) of magnesium stearate. The PCA hinted at the prevalence of impeller speed effect on ibuprofen concentration due to path length variation of the NIR light caused by the fluidized behaviour in the powder blend as a result of high speed ranges (>300 rpm). The ASCA model indicated that while both impeller speed and total feed rate effects were statistically significant (p-value=0.004), the impeller speed was the factor that contributed the most to the spectral variance (55.5 %). The PLS regression model for the ibuprofen content resulted in a RMSECV of 1.3 % (wt/wt) and showed that impeller speed was yet again the factor that exerted the major influence on spectral variance, owing to its wavelength-dependent effect that prevents common pre-processing techniques from eliminating it across the entire NIR range. The best sample presentation to the NIR probe was achieved at low impeller speed ranges (<600 rpm) and low total feed rates (<15 kg/h), such that it enhanced the PLS model ability to predict the ibuprofen concentration in the blend.

Relationship between Silicon Percentage in Graphite Anode to Achieve High-Energy-Density Lithium-Ion Batteries.

High-weight-percentage silicon (Si) in graphite (Gr) anodes face commercialization hurdles due to fundamental and interrelated challenges. Nevertheless, using the existing manufacturing line, the optimized Si/Gr ratio is the most efficient and valuable way to fabricate high-energy-density lithium-ion batteries (LIBs). Still, literature has not thoroughly examined the Si/Gr ratio. This study addresses this critical gap by systematically evaluating Si content (5-20 wt %) in commercial graphite. The goal is to optimize the Si/Gr ratio for exceptional specific capacity while mitigating inherent Si limitations like cyclic stability and first-cycle irreversible capacity loss. This work employs a multidirectional approach, including in situ electrochemical impedance spectroscopy for interface analysis, rate capability assessment (up to 3 C-rate), Li diffusion coefficient measurement, and thorough cyclic stability evaluation. Increasing the silicon (Si) weight percent from 10% to 15% in the Si15Gr75 composite anode resulted in significant improvements in the first lithiation and delithiation capacities by approximately 16.8% and 16.0%, respectively. The Si15Gr75 cell delivered a high initial Coulombic efficiency of roughly 82.9%, nearly equivalent to a pure graphite anode. Furthermore, the Si15Gr75 Li cell exhibited excellent cyclic stability at a current rate of 0.5 C, retaining about 60% of its capacity after 215 cycles. Additionally, full-cell testing against a commercial NMC622 cathode showcases excellent performance across various current rates (0.1-0.5 C). This study paves the way for the development of high-energy-density LIBs by providing valuable insights into the optimization of Si/Gr composite anodes for commercial viability.

Disentangling the socio-technical impacts of digitalization: What changes for shop-floor decision-makers?

With the diffusion of Industry 4.0 technologies, firms can decentralize operational decisions, fostering a data-driven Digital Transformation (DT) across all organizational levels. Digitalized shopfloors can leverage an unprecedented availability of data for better and faster decision-making, resulting in enhanced operational performance. However, limited research has investigated the organizational and individual implications for those who run the manufacturing lines: production managers, supervisors, team leaders, and workers. By adopting a socio-technical framework, this study aims to disentangle the effects that digitalization has on shopfloors’ organizational structures, decision-making processes, and individual competencies, as well as the interdependencies among them. An exploratory approach was adopted, based on an empirical cross-country study involving 34 semi-structured interviews conducted in the Italian and Spanish automotive sectors. Analyzed through the lenses of information-processing, knowledge-based, and dynamic capability theories, our findings reveal through five propositions how digitalization induces a “polarization” of operational decision-making: shopfloors are run by knowledgeable data-empowered production managers and autonomous information-processing team leaders on the front line, with a reduced importance of supervisors. Upskilling needs appear for team leaders but not production workers, whose involvement, however, emerges as a key factor for a successful digitalization and overall performance in initial DT stages. This study contributes to literature on digitalization by exposing managerial tensions and dynamic capabilities, along with a deeper understanding of the micro-foundations of DT in terms of implications for shop-floor decision-makers. Managerial implications are directed at creating awareness about the centrality of production team leaders, and future research avenues are proposed.

(Invited) Empowering Clean Energy Manufacturers with Low-Temperature Plasma: A Novel Approach to Real-Time Thin Film Metrology

Thin films, or micro- and nano-scale materials with unique functional properties, enable many clean energy industries, such as lithium-ion batteries, solar cells, and carbon conversion. Minimal-waste manufacturing of high-performance thin films will be vital for achieving society’s net-zero carbon emission goals. Because existing thin film sensors cannot operate within most sections of manufacturing lines, manufacturers often complete hundreds of processing steps before they can test (a small portion of) their products and detect problems. Current thin-film manufacturing thus often yields many low-quality products, or off-spec products that must be discarded. A real-time, in-situ metrology strategy can create unprecedented opportunities for implementation of smart manufacturing practices and advanced process control solutions to realize higher thin film throughput and performance.SirenOpt has developed a real-time low-temperature plasma-based thin film metrology sensor. The sensor is compact, can be integrated within existing manufacturing lines, and can collect multiple thin film property measurements (such as thickness with a precision of 1/10th of a nanometer, density, resistivity, chemical composition, and contaminate identification) in parallel and in real-time, thus allowing for the collection of critical manufacturing data that is otherwise unobtainable. SirenOpt is currently applying the sensor to industrial thin film coating processes such as Lithium-ion battery electrodes, where this flexible sensor can be used to accelerate R&D and process optimization, improve quality control, and enable real-time process control.

Evaluation of an operating durian shell charcoal briquette manufacturing line and development of a biorefinery process for higher value products

In assessing the feasibility of carbonization for the treatment of durian shell, a large-scale (10 T) fuel briquette manufacturing line is evaluated, taking into consideration the actual processing time and the effect of the biomass high moisture content on overall efficiency. The fuel product exhibits a high higher heating value (HHV) (33.21 MJ/kg) but with a low production efficiency (10 %) and a high cost of production (0.55 USD/kg briquette). One alternate biorefinery process was proposed to turn durian shells into much higher value materials, including high-methoxyl pectin (DE > 90 %), high surface area activated carbon (>300 m2/g) and mineral-rich biochar, with higher recovery efficiency and selling prices, most notably pectin at 40 USD/kg. The results from this study are then used to provide recommendations for feasible treatment, not only for durian shells but also for a variety of different complex and highly recalcitrant biomass in pursuit of a more sustainable agricultural production.

Cleaning performance of stainless steel surfaces fouled with preheated starch-protein binary mixtures: Mechanical properties, removal mechanisms, and in-place cleaning

In the food industry, deposits are often composed of multiple components and micro-structures for which removal may diverge from conventional cleaning protocols. Incomplete removal of those deposits may compromise hygiene of manufacturing lines and impact product quality. Here, we investigated the cleaning performance of model foulants made from mixtures of starch and protein adhered to stainless steel surfaces, assessing their viscoelastic properties, removal mechanisms, and cleaning effectiveness in response to different standard chemical treatments (pH 7 and pH 13) and cleaning temperatures (20 and 40 °C). Deposits displayed distinct viscoelastic behaviours under mechanical stress, affecting their removal mechanisms, especially during in-place cleaning. Young's modulus data correlated with the cleaning efficiency of the model foulants. Notably, a decrease in deposit hardness was associated with easier detachment from the metal surface for starch-rich deposits (P0 and P30, 100% and 70% starch gel respectively). In contrast, protein-rich deposits, particularly P80 (80% protein gel-20% starch gel), required greater removal forces. This was especially evident in the absence of chemical treatment, where P80 demanded more effort to be removed compared to all other deposits. The use of chemical treatment reduced the mechanical stress and removal work needed for cleaning, alkaline treatment being effective for most deposits. Alkaline cleaning was efficient at removing protein and starch-based foulants, especially for the sole protein or sole starch-containing deposits. However, P80 exhibited similar removal levels at both pH 7 and pH 13. Therefore, this research underscores the intricate removal mechanisms for starch-protein mixtures compared to single deposits, highlighting how variations in deposit composition over time during industrial processing can impact the efficiency of current Clean-in-Place (CIP) protocols.

Analysis of reject pinhole hardfacing process for stellite alloy metal motorcycle engine valves

If businesses wish to compete in the industrial world, they must produce quality and process results. No matter how big or little, some goods will always fail to generate or be rejected in a production process. In this case, companies in the automobile sector have noticed an increase in the number of pinholes that have been returned during the past three months. The company established a maximum reject rate of 0.9% each month based on the rejected data; from May to July, the average reject rate was 0.23%. Finding the factors causing the surge in pinhole rejects that occurred between May and July is the aim of this investigation. The analysis showed that the increase in pinhole rejects was caused by a variety of variables, including machine, process, and human factors. Further examination disclosed alterations to the device's configuration and software, along with the operator's persistent incapacity to comprehend the process. As operations become less efficient, repair components build up in the manufacturing line area. As a result, an improvement plan is needed, which is put into action by giving operator training and rearranging the machine and its software according to process requirements. Operators with a greater understanding of the process yield better results and higher productivity. As a result, operator confidence will rise along with performance. It was discovered that after developments in the field of human resource development, the number of rejections had dropped. Before the repair was finished, the average reject from May to July 2023 was 0.23%; following the repair, the average reject from August to October 2023 was 0.09%

Enhancing Manufacturing Excellence with Digital-Twin-Enabled Operational Monitoring and Intelligent Scheduling

This research examines the potential of digital twin (DT) technology for reformation within China’s traditional solid-wood-panel processing industry, which currently suffers from production inefficiencies and the slow adoption of digital technology. The research centers around developing a digital twin system, elucidating improvements in manufacturing efficiency, waste management, process simulation, and real-time monitoring. These capabilities facilitate immediate problem solving and offer transparency in the process. The digital twin system is comprised of physical, transport, virtual, and application layers, employing a MySQL database and using the Open Platform Communications Unified Architecture (OPC UA) protocol for communication. The application of this system has led to heightened production efficiency and better material use in the solid-wood-panel manufacturing line. Integrating the dynamic selection adaptive genetic algorithm (DSAGA) into the virtual layer drives the system’s efficiency forward. This evolved approach has allowed for an enhancement of 8.93% in the scheduling efficiency of DSAGA compared to traditional genetic algorithms (GAs), thereby contributing to increased system productivity. Real-time mapping and an advanced simulation interface have strengthened the system’s monitoring aspect. These additions enrich data visualization, leading to better comprehension and a holistic process view. This research has ignited improvements in solid-wood-panel production, illustrating the tangible benefits and representing progress in incorporating digital technology into traditional industries. This research sets a path for transforming these industries into smart manufacturing by effectively bridging the gap between physical production and digital monitoring. Furthermore, the adjustability of this approach extends beyond solid-wood-panel production, indicating the capability to expedite movement towards intelligent production in various other manufacturing sectors.

Cyclic multi-hoist scheduling with fuzzy processing times in flexible manufacturing lines

This work addresses the problem of scheduling multiple hoists (robots), which arises on real automated electroplating lines. Several hoists operate on a shared track and should not collide with each other. The processing time for each operation is preset in interval form. The problem simultaneously takes into account two criteria: maximum processing quality and maximum line productivity. To do this, we estimate the dependence of the duration of operations on the level of their quality and, in accordance with these estimates, we represent the processing times in the form of fuzzy numbers with triangular membership functions. The novelty of this study is two-fold. First, we propose a new mathematical model based on the use of fuzzy numbers and fuzzy operations on them, which is designed to deal with complex galvanic lines served by multiple hoists. The second new feature is a new combinatorial method developed for scheduling multiple hoists. For this, we modify and refine the so-called “method of prohibited intervals”, which has been actively used in scheduling theory in recent decades to solve problems, but until now was known only for the case of a single hoist. Thus, the second novelty of this work is that it is actually the first attempt to use the powerful technique of fuzzy set theory to solve a practical problem of scheduling multiple hoists. Computational experiments on real and randomly generated instances show that the new algorithm can successfully solve real problems that arise in practice.

Pharmaceutical capacity expansion under uncertainty: Framework and models

This work presents a methodology for capacity planning under uncertainty in general multi-stage manufacturing networks. Multiple manufacturing lines are available and the production time on each line must be divided between a set of materials. The methodology generates mixed-integer linear programming models that represent capacity expansions with economy of scale costs functions and production planning details. A deterministic model is solved to create a baseline and the impact of uncertainty is investigated by sensitivity analysis and stochastic programming. The applicability of the methodology is exemplified through two case studies derived from industrial pharmaceutical manufacturing. The methodology identifies bottlenecks that limit supply and, where required, activates, and assigns capacity expansion projects for satisfying demand subject to uncertainty. The methodology determines the best use of existing resources and the location and size of capacity expansions thereby generating a portfolio of recommendations for decision making on integrated planning of capacity and production.

Line Balancing Study Using Value Stream Mapping Tool on Lean Manufacturing: A Case Study in an Electronic Industry

XY Electronics, a leading international company in electronic components manufacturing, is confronting significant production constraints that adversely affect output, lead times, and operational expenses. This study examines the manufacturing line for product A using Value Stream Mapping to analyze process times and identify bottlenecks where takt times are exceeded. It focuses on areas surpassing production cycle times and aims to enhance line utilization through better line balancing and waste reduction. The results reveal that the header assembly, along with coplanarity and pre-testing 3, are major bottlenecks, which significantly impact productivity. By optimizing task allocation, refining workforce distribution, and employing cross-training, the production line efficiency improved significantly. In addition, strategic workforce reallocation and station optimization were crucial in addressing resource underutilization and enhancing overall operational efficiency

Hybrid Model Based Autonomous System for Controlling Double Side Grinding Process

Manufacturing uncertainty is rising due to diverse customer demands and new technologies. These challenges are propelling us into the Industry 4.U (Industrial Revolution 4.0 Upgrade) and Industry 5.0 era, focusing on human-centric manufacturing and value creation, necessitating digital transformation. This transformation relies on the core technologies of the 4th Industrial Revolution (Data Analytics, Connectivity, Autonomy, and Collaboration) to perform value-creating processes autonomously and reliably. To adapt, Industry 5.0-oriented Smart Manufacturing architecture was designed. In this system, process planners collaborate with generative AI to generate optimal process plans based on new specifications, pursuing human-centered manufacturing. The manufacturing system is reconfigured with modular technology, and each process is autonomously controlled as a 4.U/5.0-oriented production system. This paper focuses on developing an autonomous process control system for Industry 4.U/5.0-oriented Smart Manufacturing. The study aims to stabilize the quality of the brake disc grinding process in an automotive parts company. The developed control system has been tested on a car’s brake disc manufacturing line, with experimental results showing the reliability of the DT and AI in controlling the grinding process.

Real-time release testing of in vitro dissolution and blend uniformity in a continuous powder blending process by NIR spectroscopy and machine vision

Continuous manufacturing is gaining increasing interest in the pharmaceutical industry, also requiring real-time and non-destructive quality monitoring. Multiple studies have already addressed the possibility of surrogate in vitro dissolution testing, but the utilization has rarely been demonstrated in real-time. Therefore, in this work, the in-line applicability of an artificial intelligence-based dissolution surrogate model is developed the first time. NIR spectroscopy-based partial least squares regression and artificial neural networks were developed and tested in-line and at-line to assess the blend uniformity and dissolution of encapsulated acetylsalicylic acid (ASA) – microcrystalline cellulose (MCC) powder blends in a continuous blending process. The studied blend is related to a previously published end-to-end manufacturing line, where the varying size of the ASA crystals obtained from a continuous crystallization significantly affected the dissolution of the final product. The in-line monitoring was suitable for detecting the variations in the ASA content and dissolution caused by the feeding of ASA with different particle sizes, and the at-line predictions agreed well with the measured validation dissolution curves (f2 = 80.5). The results were further validated using machine vision-based particle size analysis. Consequently, this work could contribute to the advancement of RTRT in continuous end-to-end processes.