SACM - Australia
Permanent URI for this collectionhttps://drepo.sdl.edu.sa/handle/20.500.14154/9648
Browse
1978 results
Search Results
Item Restricted Intelligent Fault Detection for Belt Conveyor Idlers Using Machine Learning(Saudi Digital Library, 2026) Alharbi, Fahad; Luo, Suhuai; Zhang, Hongyu; Chen, Zhiyong; Wheeler, CraigConveyor belt systems are essential components of modern mining, logistics, and manufacturing operations. However, faults in idlers, which are the rollers that support and guide the belt, can reduce system reliability, cause unplanned downtime, increase maintenance costs, and create safety risks. Conventional inspection methods, such as visual assessment and manual listening, are widely used to identify abnormal idler behaviour, but they are labour intensive, subjective, and difficult to apply across large conveyor systems. Acoustic monitoring offers a promising contactless alternative because developing faults often produce changes in sound before complete failure occurs. However, these fault related acoustic signatures can be subtle and may be masked by environmental and operational noise, making them difficult to interpret using traditional signal processing methods alone. Machine learning (ML) has therefore been increasingly applied to acoustic signals to automate feature extraction, fault detection, and classification. Nevertheless, existing ML approaches still face several important challenges, including the limited availability of labelled fault data, inadequate generalisation across recording conditions and sensing platforms, and insufficient modelling of the spatial and temporal characteristics of acoustic signals. This thesis addresses these challenges by developing intelligent fault detection (IFD) models that automatically analyse acoustic signals from belt conveyor idlers. The research follows a progressive methodology comprising supervised transfer learning, semi-supervised anomaly detection, task-specific spatial–temporal deep learning, and cross-domain Convolutional Neural Network (CNN)–Transformer adaptation. To support these investigations, acoustic datasets were collected using handheld microphones and drone-mounted recorders under multiple idler operating and fault conditions. The first study investigated supervised fault classification using the original handheld acoustic dataset, which contained 255 four-second samples representing Normal, Stage 1, Stage 2, and Stage 3 operating conditions. Embeddings were extracted using YAMNet, a pre-trained audio neural network, and processed using Bidirectional Long Short-Term Memory (BiLSTM) and Bidirectional Gated Recurrent Unit (BiGRU) networks to model contextual relationships in both forward and backward temporal directions. Attention mechanisms and an Extreme Gradient Boosting (XGBoost) classifier were also evaluated. The best-performing YAMNet–BiLSTM configuration achieved an accuracy of 90.59% and an F1-score of 90.57% for four-class fault-stage classification. Although this study established a strong supervised baseline, it required labelled examples from every operating condition and fault stage. This dependence limits practical application because faulty-idler recordings are relatively rare, costly to collect, and difficult to label in industrial environments. To address this limitation, the second study developed CASSAD (Chroma-Augmented Semi-Supervised Anomaly Detection), which was trained exclusively on normal operating samples during training. Following the first study, additional handheld recordings were collected, increasing the dataset from 255 to 468 acoustic samples. This larger dataset enabled CASSAD to be evaluated using a broader set of normal and abnormal operating recordings. CASSAD combines Chroma-STFT, Chroma-CQT, and Chroma-CENS representations with filtering, statistical aggregation, and a one-class support vector machine (OC-SVM). On the expanded 468-sample handheld dataset, the best-performing CASSAD configuration achieved an accuracy of 90.59%, a positive-class F1-score of 92.42%, and an Area Under the Receiver Operating Characteristic Curve of 96.29%. To enable a direct comparison with the YAMNet-based models, CASSAD was also evaluated on the original 255-sample dataset after the three fault stages had been combined into a single Abnormal class. In this binary evaluation, CASSAD achieved an accuracy of 93.00% and an F1-score of 93.25%, compared with an accuracy of 92.18% and an F1-score of 93.00% for the strongest YAMNet-based configuration. These results demonstrate that competitive anomaly-detection performance can be achieved without labelled abnormal samples during training. However, CASSAD provides only binary Normal–Abnormal decisions and uses temporally aggregated features, limiting its ability to distinguish fault severity and model changes in acoustic behaviour over time. To overcome these limitations, the third study developed TD-CLNet, a Time-Distributed CNN–Long Short-Term Memory (LSTM) architecture designed to perform multi-stage fault classification while learning spatial and temporal representations from the acoustic data. The expanded 468-sample handheld dataset was re-segmented into one-second samples and converted into log-Mel-spectrogram frames. TD-CLNet applies a shared CNN feature extractor to each frame and then uses an LSTM to model the resulting temporal feature sequence. This design enables the model to distinguish among Normal, Stage 1, Stage 2, and Stage 3 conditions while preserving temporal information that was reduced through the statistical aggregation used in CASSAD. Under four-fold cross-validation, TD-CLNet achieved a mean accuracy, precision, recall, and weighted F1-score of 92.1%. It outperformed the evaluated conventional CNN–LSTM configurations and provided a small improvement over the strongest YAMNet model re-evaluated on the same expanded dataset. Nevertheless, the sequential processing used by LSTM networks limits parallel computation and provides less direct access to broader global relationships within the acoustic feature sequence. To address these limitations and investigate cross-domain generalisation, the fourth study developed hybrid CNN–Transformer models using four pre-trained CNN backbones: ResNet-18, DenseNet-121, EfficientNet-B0, and ShuffleNet-V2. Two acoustic feature representations were investigated: Mel-spectrograms, which represent the distribution of signal energy across perceptually scaled frequency bands over time, and Mel-Frequency Cepstral Coefficients (MFCCs), which provide a compact representation of the short-term spectral envelope. The CNN backbones extracted local spectral representations, while Transformer encoders modelled broader contextual relationships within the feature sequences. In the first phase, the models were trained and evaluated using 0.5-second segments derived from the handheld source-domain recordings. The ResNet-18–Transformer models achieved a cross-fold mean accuracy of 96.9%, with a 95% confidence interval of 96.4%–97.6%, while a ResNet-18–Transformer ensemble increased the handheld-domain accuracy to 98.0%. In the second phase, the models trained on the handheld source-domain dataset were adapted to the drone-acquired target-domain dataset. The drone recordings represented a more challenging sensing environment because they were affected by rotor noise, changing recording distances, varying microphone positions, and environmental interference. Multiple fine-tuning strategies were evaluated, included full fine-tuning (Full-FT), freezing the CNN backbone (CNN-Frozen), freezing the Transformer encoder (TR-Frozen), and 𝐿2-SP regularisation. Mean–Covariance Alignment (MCA) was compared with a cross-entropy (CE) baseline and several established domain-adaptation methods. MFCCs produced the strongest CE baseline in several drone-domain experiments, achieving an accuracy of 90.5% under 𝐿2-SP. However, MFCC performance generally decreased when MCA was applied. In contrast, MCA improved the Mel-spectrogram Full-FT configuration, increasing accuracy from 84.7% to 86.0% and the F1-score from 83.4% to 85.6%. These findings demonstrate that domain-adaptation performance depends on the interaction among the acoustic representation, fine-tuning strategy, model architecture, and alignment objective; no single adaptation method was uniformly optimal across all evaluated configurations. In summary, this thesis establishes a connected research pathway for acoustic fault detection in belt conveyor idlers. It provides handheld and drone-acquired acoustic datasets, establishes supervised transfer-learning baselines, introduces CASSAD for anomaly detection without labelled abnormal training samples, develops TD-CLNet for task-specific spatial–temporal fault-stage classification, and proposes a two-phase CNN–Transformer framework for handheld-to-drone domain adaptation. Collectively, these contributions advance acoustic idler monitoring towards more accurate, data-efficient, and adaptable fault detection while identifying the need for further validation across additional industrial sites, conveyor configurations, sensing platforms, and operating conditions before large-scale real-world deployment.15 0Item Restricted Multiscaling Asymptotic Behaviour of Random Fields Driven by SPDEs(Saudi Digital Library, 2026) Alghamdi, Maha Mosaad A; Andriy, Olenko; Nikolai, LeonenkoThis thesis studies the asymptotic behaviour of random fields which are solutions of stochastic partial differential equations with random initial conditions exhibiting long-range and cyclic long-range dependence. The main focus is on investigating multiscaling limits and their spectral and covariance structures. The cases of solutions to fractional, high-order, and Riesz–Bessel type equations are considered. The results are obtained for the cases of arbitrary multiple spectral singularities, which have not been considered before. First, the thesis investigates multiscaling limits for stochastic fractional equations with random initial conditions possessing cyclic long-memory. The corresponding rescaled solutions are shown to converge to Gaussian limit fields. Their explicit spectral and covariance representations are provided. Next, the thesis studies high-order heat equations with random initial conditions whose spectra have singularities at zero and at non-zero frequencies. Using spectral and scaling techniques, it is proved that the normalised solutions converge to Gaussian random fields determined by the presence or absence of a zerofrequency singularity. Kernel averaging is introduced for odd-order equations to obtain non-degenerate limits. Finally, the thesis analyses fractional Riesz–Bessel equations with initial conditions exhibiting both classical and cyclic long-range dependence. It is proved that the rescaled solutions converge to spatio-temporal Gaussian random fields that are stationary in space and non-stationary in time. In addition, multiscaling limit theorems are derived for the case of regularly varying asymptotics.4 0Item Restricted The Influence of Saudi Women’s Empowerment on Female Characters in Children’s Literature: A Feminist Critical Discourse Analysis(Saudi Digital Library, 2026) Alomari, Shatha; Bulfin, Scottتكمن أهمية هذه الدراسة في محاولتها فهم التقاطع بين أدب الأطفال والواقع الاجتماعي، إذ تستكشف كيف أثّرت سياسات تمكين المرأة في السعودية على الشخصيات النساء وتمثيلها في قصص الأطفال من خلال مقارنة أربعة قصص: اثنتان نُشرتا قبل عام 2016 واثنتان بعده. تناولت الدراسة ثلاثة أسئلة رئيسية: الأدوار التي أُسنِدت للشخصيات النسائية، ومستوى الوكالة التي مارستها داخل السرد، ومدى اتساق القصص المنشورة بعد مرحلة التمكين مع الخطاب الوطني حول تمكين المرأة. اعتمد البحث منهجًا نوعيا مقارنًا يستند إلى التحليل النسوي النقدي للخطاب، مستخدما نموذج فيركلوف الثلاثي كأداة التحليل الأساسية. أظهرت النتائج استمرار الأدوار الجندرية التقليدية في القصص الأربع، في حين أبدت الشخصيات النسائية درجات متفاوتة من الوكالة تأثرت بالسياق الاجتماعي لكل فترة زمنية. كما كشف التحليل أن القصص المنشورة بعد التمكين عكست الخطاب الإعلامي المصاحب لرؤية 2030، وفي الوقت نفسه أعادت إنتاج بعض الأيديولوجيات الجندرية وقاومت بعضها الآخر. وبناءً على هذه النتائج، يمكن القول إن القصص اللاحقة لمرحلة التمكين تأثرت بسياسات تمكين المرأة. ومع ذلك، تبقى النتائج محدودة بسبب صغر حجم العينة، وهو أحد قيود الدراسة. لكن تفتح هذه الدراسة المجال أمام مزيد من الأبحاث حول تأثر أدب الأطفال بالتحولات الاجتماعية.20 0Item Restricted Identifying Cooler and Flood-Safer Urban Zones on Riyadh's Fringe Using a Digital Elevation Model (DEM) and the National Geoid Model KSA-Geoid21(Saudi Digital Library, 2026) Aljebreen, Suliman; Dewan, AshrafThis thesis develops a GIS-based suitability framework to identify relatively cooler and flood-safer zones for urban fringe planning in Riyadh, Saudi Arabia. The study integrates two physical indicators: (i) seasonal Land Surface Temperature (LST) derived from Landsat data, and (ii) DEM-derived hydrological indicators including D-Infinity flow routing, wadi channel delineation, flood-exposure modelling, and terrain slope. These indicators were standardised and combined using Weighted Linear Combination (WLC) under three weighting scenarios. A ±15% sensitivity analysis using Spearman’s rank correlation was used to assess the stability of the resulting suitability patterns. The results show that most of Riyadh’s fringe remains constrained by high thermal stress and/or flood-related exposure, while comparatively more suitable zones are concentrated mainly in the western and south-western fringe. The identified spatial pattern remained reasonably stable under different weighting scenarios, suggesting that the framework can support comparative planning decisions. All analyses were undertaken in UTM Zone 38N (EPSG:32638). The outputs include suitability maps, extracted priority zones, and summary statistics relevant to physically informed urban fringe planning in Riyadh.3 0Item Restricted GOBEE - Gamified Outdoor Balanced Eating and Exercise(Saudi Digital Library, 2026) Alsaeed, Hanan; mary, Jabbarzadeh; Harry, WangChildhood obesity remains one of Australia's most significant preventable public health challenges, with 27.7% of children classified as overweight or obese and fewer than 9% meeting recommended daily vegetable intake. Gobee is a mobile-first Progressive Web Application designed to address this gap by combining gamified outdoor activity with AI-assisted healthy meal discovery for families with children aged 5–12. The platform integrates two core features — City Quest, which converts GPS-guided neighbourhood walks into reward-based activity challenges, and Chef Cam, which uses computer vision (OpenAI GPT-4o) to detect fridge and pantry ingredients and recommend healthy, child-friendly recipes from a database of over 700,000 items. Both features feed into a virtual pet progression system that provides children with intrinsic motivation through visible, positive-reinforcement-based rewards, while requiring no social interaction, public leaderboards, or user-generated content, preserving child safety by design. Architecturally, Gobee is built on a modular, serverless cloud infrastructure hosted on AWS (Lambda, API Gateway, Cognito, RDS, S3, and Amplify), with a Vue.js/TypeScript frontend and a FastAPI backend connected to separate PostgreSQL and MySQL databases for recipe and gameplay data respectively. The system was engineered with maintainability, scalability, and compliance as first-order design constraints: it aligns with the Australian Privacy Principles under the Privacy Act 1988 (Cth), the OWASP Top 10 (2021) web security guidelines, and the ACS Code of Ethics, minimising personal data collection and ensuring all AI-generated recommendations remain advisory and subject to parental review. Beyond its technical implementation, Gobee is positioned as a scalable, low-distribution-cost platform for potential sponsorship by government health departments, health insurers, and school networks seeking evidence-based digital tools for preventative family health.6 0Item Restricted Understanding Alloy Performance Trends in Real Ores(Saudi Digital Library, 2026) AlMahfood, Mahdi; Ali, Yahia; Gates, JeffResearch Problem and Context In heavy mineral processing infrastructure, ore transfer chutes are subjected to relentless abrasive sliding and low-angle kinetic impacts by high-tonnage particulate product streams. Sacrificial protective liners, typically specified from homogeneous quenched-and-tempered steels or multi phase high-chromium white cast irons (HCWCIs), display rapid deterioration rates in the field. Materials specifiers and mechanical asset engineers traditionally pre-screen candidate alloys using standard laboratory tests operating with benchmark quarry aggregates (such as basalt or granite) to forecast wear life. However, a systemic disparity exists between the optimistic alloy performance profiles predicted by standard laboratory models and the accelerated, premature wear rates documented in operating mining installations processing real ores. This contradiction stems from an unquantified limitation: standard quarry minerals do not accurately replicate the unique physical characteristics, high specific gravities, and brittle fragmentation transitions of true multi-component mining rocks. Consequently, relying on quarry rock proxies risks generating misleading alloy selection hierarchies, resulting in unexpected plant structural failures and costly unscheduled maintenance shutdowns. Methodology Overview This investigation evaluated the mechanical and metallurgical validity of laboratory quarry proxies by systematically comparing the wear rates and performance rankings of an established six-material alloy suite across both industrial and quarry minerals. Testing was executed using the Inner Circumference Abrasion Test (ICAT) apparatus operating in a sliding and low-angle impingement configuration (10-degree specimen tilt angle) at a controlled paddle tip velocity of 9 m/s, utilising a tightly sieved coarse fraction sizing of -8 mm +1 mm. The experimental matrix evaluated four geologically distinct rock types: standard commercial Basalt and Granite (quarry benchmarks), alongside real, high-competence Marandoo Iron Ore and heterogeneous Oyu Tolgoi Quartz Monzo diorite (QMD) ore (industrial benchmarks). Quantitative mass loss measurements were taken before and after each run to map cumulative mass to-volume loss and linear thickness reduction tracks (micrometres per hour) across five replicates. Following each primary test, a specialised dual-reinforced high-carbide white iron specimen (CB102) was isolated for microstructural comparative analysis. Qualitative surface diagnostics were conducted via high-vacuum Scanning Electron Microscopy (SEM) using a Hitachi TM3030. To isolate fundamental micro-mechanisms, four subsequent Very Short Duration (VSD) scratch mapping experiments were conducted using highly polished CB102 specimens across all four abrasive ore types to characterise immediate steady-state profile changes. Key Findings and Results The quantitative wear data revealed that the industrial ores were substantially more severe than their assumed quarry proxies, with QMD producing higher wear rates than granite for metallic materials, and iron ore producing higher wear rates than basalt across all tested materials. Material optimisation tracking demonstrated that for highly abrasive QMD and granite environments, sintered technical structural ceramic (CE45, 92% alumina) provided superior wear resistance. Conversely, in iron ore and basalt applications, the specialised micro-alloyed white iron (CB102) emerged as the most suitable material. SEM microstructural diagnostics elucidated the precise wear mechanisms driving these variations. In QMD environments, high-hardness quartz phases sheared directly through the matrix and the (𝐶𝑟,𝐹𝑒)7𝐶 eutectic carbides as if they offered equal resistance, though the harder Niobium Carbides (NbC) successfully resisted abrasion and protruded from the surface. Both carbide phases sustained extensive localised micro-cracking without complete fragmentation or pull-out. In contrast, basalt minerals lacked the hardness to cut the eutectic carbides, eroding only the softer matrix and leaving the flat carbide plateaus heavily recessed. Testing with iron ore revealed an intensely adhesive, "sticky" interface, masking over 95% of the metallic surface area. However, visible regions indicated an abrasive mechanism closely mirroring that of basalt, albeit with a less pronounced height contrast between the matrix and eutectic phases. Implications and Significance These results clearly demonstrate that standard construction quarry aggregates are fundamentally unsuitable standalone proxies for industrial mining ores due to severe misalignments in micro abrasive mechanics. Relying on standard testing without establishing strict mathematical correlation factors risks underestimating liner wear severity and may lead to poor material selection in mining infrastructure. For asset managers, this work establishes a clear material selection hierarchy: specifying structural ceramics to combat quartz-heavy gouging (QMD) and deploying micro alloyed white irons (NbC reinforced) to mitigate highly adhesive iron ore wear streams.21 0Item Restricted Examining the role of mRNA-LNP components in modulating inflammatory responses to mRNA vaccines(Saudi Digital Library, 2026) Algarni, Azizah Ali S; Angus, Johnston; Colin, PoutonCurrent licensed mRNA vaccines have been reported to cause inflammatory side effects in some vaccinated individuals. The main objective of this work was to investigate correlations between mRNA and lipid nanoparticles (LNP) components and systemic inflammatory responses. Our findings revealed strong associations between ionizable lipid structure, mRNA chemical modifications, and the magnitude and profile of systemic inflammation following exposure to mRNA–LNP formulations. Collectively, these results highlight the importance of understanding the molecular features of mRNA–LNP components to guide the rational design of safer and more effective mRNA vaccines20 0Item Restricted From Platform Logic to Creative Agency: Negotiating Visibility on TikTok and Instagram in the Saudi Context(Saudi Digital Library, 2026) Abdulsamad, Abdulrahman; Berry, Marsha; Burke, Elizabeth; Peterson, Renéeيستكشف البحث دور منصات التواصل الاجتماعي بوصفها بيئات ديناميكية تُشكّل الإدراك البشري والظهور الرقمي. وتُسهم التحيزات التي تفرضها المنصات، إلى جانب الفهم المحدود لدى منشئي المحتوى لأنظمة المنصات (Souza, 2020)، في تحديات الظهور ضمن مشهد وسائل التواصل الاجتماعي في المملكة العربية السعودية. ويجادل هذا البحث بأن التفاوض الإبداعي يوفّر وسيلة للتعامل مع هذه القيود دون المساس بالقيمة المهنية. تجمع المنهجية بين الأدبيات البحثية القائمة على البحث حول ديناميكيات المنصات وأنظمة التوصية، والبحث القائم على الممارسة من خلال إنتاج ثلاثة مقاطع فيديو قصيرة، إلى جانب الممارسة التأملية لتقييم قرارات الإنتاج. وقد أُنتج كل مقطع فيديو بنسختين، مع مقارنة الأداء من خلال عدد المشاهدات. وتُظهر النتائج أن جودة الإنتاج، والتحول إلى التنسيق الرأسي (9:16)، والتوقيت الاستراتيجي، واستخدام الوسوم (#hashtag)، تؤدي دور أدوات معرفية تترجم السلوك البشري إلى أولويات الظهور التي تعتمدها المنصات. ويقدّم البحث نموذجًا تطبيقيًا يربط بين منطق المنصات والفاعلية الإبداعية، ويطرح استراتيجيات عملية لتحسين الوصول إلى الجمهور.14 0Item Restricted Spatiotemporal Analysis of Daytime and Nighttime Urban Heatwave Dynamics: A Comparative Study of Riyadh and Perth (2003–2024)(Saudi Digital Library, 2026) Alqahtani, Mohammed Hadi S; Ashraf, DewanThis report examines the daytime and nighttime urban heatwave dynamics in Riyadh, Saudi Arabia (hot desert, BWh) and Perth, Western Australia (Mediterranean, Csa) from 2003 to 2024. The daily maximum (Tmax) and minimum (Tmin) temperatures were retrieved from the ERA5-Land hourly reanalysis data within the urban boundaries of each city. Heatwaves were defined as exceeding the 90th percentile of Tmax and Tmin based on an 11-year baseline period (2003–2013), and must include at least three days of exceedance. The frequency, duration and intensity of the heatwaves were computed annually for both day and night time events and long-term trends were evaluated by Mann–Kendall and Theil–Sen slope estimator. The frequency of heatwaves is statistically significant increasing in both cities between 2003 and 2024. The frequency trend during the day is slightly more pronounced than during the night, but both are highly significant, as would be expected in an arid climate with low atmospheric humidity which means that overnight cooling is limited, which leads to amplification of heat extremes. The sea breeze effect of the Fremantle Doctor is also evident in Perth, with the magnitude of the trend similar but of lesser magnitude than that of Riyadh during day and night. Spatially, heatwave hotspots are more intense in the dense central districts of Riyadh, and lower intensity along the coast and higher intensity in the eastern suburbs (inland) of Perth. These results show that both cities are experiencing increased heat exposure and offer a climatological baseline at the urban scale for adaptation planning to heat. These findings carry direct implications for urban heat adaptation: the near similarity of daytime and nighttime heatwave trends in Riyadh signals limited nocturnal recovery for outdoor and informal workers, while the persistence of a warming trend in Perth despite the moderating Fremantle Doctor suggests that coastal sea-breeze buffering may not be sufficient to offset background warming in the longer term. The spatial intensity patterns identified at the 9 km grid scale provide an evidence base for prioritising heat-adaptation planning (e.g., shading, cool-roof, and urban greening programmes) toward the central Riyadh and inland eastern Perth districts identified as hotspots, although finer-resolution data would be required to target interventions at the street or building scale.4 0Item Restricted Shelf-Life Assessment of Sauces for Commercial Release(Saudi Digital Library, 2026) Ahmed, Rahaf Mohammed M; Alsahafi, Naelah Abdulali A; Penny, Brotja; Jayani, Chandrapalaتهدف هذه الدراسة إلى تقييم الاستقرار الفيزيائي الكيميائي، والميكروبيولوجي، والحسي لثلاث تركيبات جديدة من الصلصات التجارية لتأسيس فترة صلاحية تجارية مبنية على مبررات علمية دقيقة. خضعت العينات لظروف تخزين مبردة 4°C، وفي درجة حرارة الغرفة 22°C، ومسرّعة 35°C خلال فترة التخزين. طوال فترة الدراسة، تمت مراقبة الاستقرار الفيزيائي الكيميائي عبر قياس درجة الحموضة pH، والنشاط المائي aw، وإجمالي المواد الصلبة الذائبة °Brix، واللزوجة، وفرق اللون الكلي ΔE. كما تم التحقق من السلامة الميكروبيولوجية باستخدام فحص إجمالي العدد البكتيري TPC، وبكتيريا الإشريكية القولونية والقولونيات، وتعداد الخمائر والفطريات. ولتحديد مدى القبول العملي للمنتج، قامت لجنة مستهلكين داخلية بتقييم الخصائص الحسية التي شملت اللون، واللمعان، والرائحة، والقوام، وانفصال السوائل. إثر ذلك، طُبقت نمذجة أرهينيوس الحركية Arrhenius على بيانات التغير اللوني لتقدير العمر الافتراضي التجاري. أظهرت النتائج أن البيئة شديدة الحموضة للصلصات حدّت من النمو الميكروبي، مما عزز الاستقرار الميكروبيولوجي للمنتجات وأشار إلى عدم وجود نمو ميكروبي قابل للكشف خلال فترة التخزين المدروسة. ومع ذلك، برز تعارض جوهري بين التنبؤات الآلية والبيانات الحسية الواقعية؛ ففي حين تنبأ نموذج أرهينيوس بفترة صلاحية ممتدة عند درجة حرارة الغرفة 22°C، أظهرت التقييمات الحسية رفض العينات المخزنة عند درجة حرارة الغرفة والظروف المسرّعة 35°C في مراحل مبكرة. ويُعزى هذا الرفض بشكل أساسي إلى الانفصال الفيزيائي للسوائل Syneresis والانهيار الهيكلي للقوام، والذي حدث قبل وصول تغيرات اللون إلى الحدود الحرجة المحددة رياضيًا. ومن بين التركيبات المدروسة، أظهرت صلصة الباربكيو BBQ أعلى مستوى من الاستقرار الهيكلي، بينما أظهرت صلصة التتبيلة Dressing حساسية أعلى للإجهاد الحراري. تَخْلُص الدراسة إلى أن الاعتماد حصرياً على الحدود الرياضية للون باستخدام نموذج أرهينيوس يُعد غير كافٍ لتحديد العمر الافتراضي التجاري للصلصات، إذ كان التدهور الفيزيائي والقبول الحسي العاملين الأكثر تأثيراً في تحديد فترة الصلاحية الفعلية. وتشير النتائج إلى أن ظروف التخزين المبردة عند 4°C حافظت على جودة المنتجات واستقرار خصائصها الفيزيائية والحسية، مما يدعم اعتماد التبريد كشرط أساسي للحفاظ على جودة الصلصات وقبول المستهلك.23 0