Renewable energy sources are seen as the key to future energy needs because they are clean and sustainable. Solar photovoltaic (PV) energy is a widely used and accessible renewable source. This research introduces a new way to enhance the efficiency and reliability of PV power systems. The study focuses on using high-performance, grid-tied hybrid DC-DC converters in PV systems. These converters can minimize power loss during energy conversion. Various well-known existing MPPT (Maximum Power Point Tracking) algorithms are used to obtain maximum power output from the PV modules. However, owing to their operational characteristics, it takes them longer to track MPP. Additionally, they provide oscillation near MPP. An improved MPPT algorithm based on the pelican optimization method is developed to ensure optimal power from the PV source by reducing tracking time and minimizing fluctuations around MPP. The research also discusses the causes of power loss, including those caused by parasitic resistance in components. It then proposes design strategies to reduce these losses. Mathematical simulations show that the proposed system attains high efficiency across a wide range of power levels. The design also minimizes the impact on voltage gain and voltage stress caused by load and parasitic resistance changes. Theoretical models are developed for component resistances, efficiency, voltage gain, and voltage stress. To address challenges like the intermittent output and irregular voltage of PV power generation, this research presents a system that combines a high-performance DC voltage conditioner with an efficient inverter. A prototype system is built and tested employing simulations (MATLAB/Simulink) and theoretical calculations to verify the effectiveness of the proposed design. This method effectively maximizes energy production and ensures compatibility with the power grid by maintaining low Total Harmonic Distortion (THD) and high power factor. Additionally, the system optimizes power conversion at each stage through advanced mathematical modeling. The findings from this research, including the proposed system design and data results, are valuable for professionals and researchers in renewable energy and battery management systems (BMS) for electric vehicles. This paves the way for further advancements in these fields.

Hydrogen-based systems are essential for transitioning to a sustainable and low-carbon future. They provide a clean, efficient, and versatile energy solution by utilizing hydrogen as a fuel, which produces only water as a byproduct when burned or used in fuel cells. Therefore, the current research explores innovative energy systems for hydrogen production and their integration into sustainable energy frameworks. Three advanced techniques for hydrogen generation are analysed, each leveraging renewable energy sources to reduce dependence on fossil fuels. The first system employs a thermochemical cycle to convert excess electrical energy into hydrogen, enabling efficient energy storage and utilization. After analysis of system 1, the results reveal that the system generates 0.36 kg/h of hydrogen, which is utilised via SOFCs to produce power and electricity. Additionally, 1002 kW of electricity and 3.6 kW of heating can be supplied via the developed system. The overall system can achieve energy and exergy efficiencies of 20.6% and 22.1%, respectively. However, the thermochemical cycle can achieve maximum energy and exergy efficiencies of 79.6% and 62.9%, respectively. The second system comprises a solar PV-driven AEM electrolyser for hydrogen production. Subsequently, the produced hydrogen is fed to the fuel cell for electricity. The system's performance based on different kinds of energy and exergy efficiencies has been examined for the Kuching and Kota Kinabalu cities. After the analysis, it is concluded that system 2 can produce a maximum hydrogen production rate of 0.29 kg/h at 0.19 kW/m2 global radiation for the Kuching city whereas, for the Kota Kinabalu city, the highest hydrogen production rate can be 0.26 kg/h at 0.25 kW/m2 global radiation. Furthermore, the system's overall energy and exergy efficiencies are evaluated as 5.1% and 5.4%, respectively. Also, the effect of current density on the hydrogen production rate is analysed, and the result suggests that the hydrogen production rate decreases with a rise in current density. The third system of the current study consists of a photoelectrochemical AEM reactor, which uses low-cost earth metal-based electrocatalysts to produce hydrogen. Like PV-assisted AEM electrolyser, the performance of the PEC AEM reactor is assessed for different operating conditions when utilized in two different cities, i.e., Kuching and Kota Kinabalu. It is concluded that system 3 can produce a maximum hydrogen rate of 0.25 kg/h in Kuching and Kota Kinabalu. Furthermore, it exhibits the same energy and exergy efficiencies for both cities, i.e., 9.1% and 9.6%, respectively. The study also finds that while the hydrogen mass flow rate increases with higher solar irradiation and illuminated areas, the overall energy and exergy efficiencies decline with larger photocathode illumination areas. After comparing the three systems, it is concluded that system 1 (thermochemical cycle-based hydrogen production) outperforms all the systems in terms of energy and exergy efficiencies. The proposed hydrogen production system, like thermochemical cycles and PEC AEM reactors, can be used for various sectors, enabling a cleaner, more sustainable energy ecosystem. The future applications can include energy storage and grid balancing, ammonia and fertilizer production, carbon capture integration, etc. The current study's results help researchers improve efficiency, reduce costs, and enhance the sustainability of the different hydrogen production and storage systems.

The prevalence of Autism Spectrum Disorder (ASD) among children in the United States was reported at 1 in 59 for children aged 8 years. In Malaysia, while official statistics are limited, a preliminary investigation carried out by the Ministry of Health Malaysia, examining children aged 18 to 26 months, revealed a prevalence rate of 1.6 per 1000 children. Numerous cases remain undiagnosed, yet the escalating number of autism cases observed by healthcare professionals in paediatric centres strongly indicates a potentially higher prevalence rate in Malaysia. Children with ASD encounter difficulties in expressing affective states, particularly due to a deficit in socio-emotional communication skills. Understanding their affective states is crucial, yet conventional assessment methods, such as EEG and ECG, which often involve the use of patches, can be invasive and may lead to emotional distress. These methods disrupt natural behaviours, leading to inaccurate representations of their affective states. Recognizing the need for a more effective approach, the research proposes non-invasive methods to assess the affective states of children with ASD. It introduces a novel framework for modelling of affective states by using Convolutional Neural Network (CNN) classifier. The research recruited 56 children as the subject, comprising 28 ASD children aged between five and nine years (M = 6.43, SD = 1.2), and an additional 28 typically developing (TD) children (M = 5.65, SD = 2.2) serving as the control group. The investigation focused on the frontal facial thermal imaging of ASD children in response to specially developed video stimuli representing five primary affective states. To ensure the accuracy and impartiality of the assessment, the stimuli were verified by expert blind coders through questionnaires. These questionnaires captured the subjects’ responses to each video stimulus, evaluating valence and arousal levels. The thermal imaging data of children with ASD exhibited unique patterns associated with cutaneous blood flow under the skin regulated by the Autonomic Nervous System (ANS). These patterns were associated with the five basic affective states when the children were exposed to the video stimuli. The research leveraged GLCM, wavelet coefficients, and thermal intensity values from specific regions of interest (ROI) in the facial image as input features for the CNN model. The model enabled real-time computation of affective state outputs, facilitating quantifiable correlations between temperature patterns and affective states. Statistical analysis evaluated these correlations in forms of valence and arousal values. The responses were then mapped onto Kollias's 2-D Circumplex Model of Affect to validate the affective state model. The proposed model is capable of classifying the affective states with high accuracy of 94.10% for TD, 89.60% for ASD, and precision of 95.76% for TD, 91.66% for ASD. The validity of the approach was further confirmed using the CK+ and Rusli et al. frontal facial databases, demonstrating notable performance with an accuracy of 91.81% and precision of 94.54%. These findings reveal the potential of using non-invasive and less intrusive method through thermal imaging with advanced machine learning techniques in assessing real-time affective states in autistic children. It facilitates a more effective diagnostic and early intervention therapy.

RGB-D based Face Recognition (FR) has grown popular with low-cost depth cameras like Microsoft Kinect, Intel RealSense, and Zed. However, these advances are insufficient for Open-World Face Recognition (OWFR) because the recognition model must identify individuals for whom the FR model was not trained. Developing robust open-world face recognition systems is critical for many practical applications such as cobots, law enforcement, security, and surveillance. Existing FR methods require extensive fine-tuning, classifier retraining, or global metric learning to improve the performance for effective domain adaptation in the open world. These steps are computationally expensive and time consuming. The recognition performance will also significantly degrade when presented with unique individuals. Therefore, it is necessary to develop robust multimodal open-world FR systems using RGB-D cameras without incurring substantial downtime. This thesis proposes three main contributions to the research in RGB-D face recognition. Firstly, the thesis investigates and proposes an RGB-D based FR models suited for the open world, termed CuteFace3D. These robust FR models are built using a multimodal CNN and RGB-D face dataset. The various CNN backbones are investigated for the task. The close-set evaluation on the Intellifusion test dataset is used as the criterion to select a more discriminative FR model as a feature extractor for OWFR. The selected models are then extensively analyzed for an open world on a large dataset of 3D faces. The results imply that deeper networks alone are not discriminative enough for OWFR. The storage is optimized by eliminating the need to save raw RGB-D images, reducing model inference time, and improving data security. A complete FR pipeline is also implemented using a RealSense D435 depth camera. In addition, embeddings are utilized for open-world and unseen domain adaptation with the KNN classifier and k-fold validation, which achieved 99.997% for the open set RGB-D pipeline for domain adaptation. Early fusion with multichannel RGB-D input makes the proposed models robust and accurate in open-world scenarios, with performance equivalent to close-set FR models. Secondly, for OWFR, a fast and efficient adaptive threshold algorithm is developed using an effective Region of Interest (ROI) setting for metric learning. It uses five different ROI schemes to find an adaptive threshold in real-time. After new enrolment the algorithm determines the FR model’s quality and usability. To establish the effectiveness, then benchmarked the proposed method against various threshold-finding strategies for face recognition algorithms for open-world adaptation on different datasets. Experimental results demonstrated that the proposed ROI-based method is up to 12 times faster than the best threshold search algorithm, reporting higher accuracy and fewer errors. Thirdly, this thesis also addresses the performance degradation of the FR model in an open-world setting. A novel performance evaluation metric for FR algorithms on imbalanced datasets is proposed. The proposed metric with an adaptive threshold is more effective than conventional fixed threshold strategies. Thus, this thesis alludes that FR algorithms should be benchmarked for accuracy at the highest F1-score in an open-world. In conclusion, all three contributions increase the effectiveness and efficiency of the proposed FR in terms of computational cost, storage, and security. The proposed method also reduces computational time, making existing FR models operational for OWFR in real-time.

Electric vehicles (EVs) have become a favourable choice due to the current environmental conditions and limited fuel resources. For efficient operation, EVs often use a lithium-ion battery as its main power source. Nevertheless, during acceleration, EVs require an instant high load demand, which is quite challenging to satisfy with the lithium-ion battery alone due to its slow discharging rate. This frequent fluctuation can damage the batteries’ State of Health (SoH), and to overcome this issue, a Hybrid Energy Storage System (HESS) is proposed. In the system, a Supercapacitor (SC) is used to support the immediate load demand from a vehicle. To ensure that the correct amount of power is extracted, a suitable controller needs to be integrated with a Bidirectional DC-DC Converter (BDC). As a model disturbance can influence both the load demand and system feedback response, a novel contribution of this work is to introduce the application of Model Reference Adaptive Control (MRAC) to overcome this issue. A detailed derivation of this algorithm, along with the investigation of the tuning effect, is presented. To analyse the efficacy of this controller, several numerical simulations have been carried out using MATLAB/Simulink, where the MRAC performance is benchmarked against the Proportional Integral (PI) controller, based on several performance indexes such as Root Mean Square Error (RMSE) of current and voltage, power demand tracking, and controllers’ characteristics. For regular operation, the results show that MRAC outperforms the PI controller in tracking voltage demand by 67% (with constant voltage) and 85% (with variable voltage) with inverting BDC and current demands by 16% (with variable current) in non-inverting BDC. While in the presence of disturbance, MRAC shows its efficacy in current demand tracking by surpassing PI controller with 15% higher accuracy. In this case, MRAC requires some time due to adjust its mechanism to surpass the PI controller in tracking the load demand. To validate the MRAC design, an EV model, designed by MathWorks has been utilised upon the integration of the HESS with a Power Management System (PMS) that operated with four (4) different driving cycles, approved by the Environmental Protection Agency (EPA), such as US06, Urban Dynamometer Driving Schedule (UDDS), Highway Fuel Economy Test (HWEFT) and Federal Test Procedure (FTP). The comparison results show MRAC consistently demonstrates superior current tracking compared to PI controller under disturbance conditions, as evidenced by significantly lower RMSE values in HWFET (8.15 vs. 39.74), UDDS (7.4 vs. 31.97), and FTP (6.34 vs. 24.89) drive cycles, respectively. Finally, the results of this study highlight the potential of adaptive control strategies in improving the efficiency, stability, and reliability of power management systems along with BDC for Hybrid Electric Vehicles (HEVs).

Due to urgent needs in daily life, our demand for highly complex and sophisticated automatic control systems is becoming inevitable. That was not possible before the advancement of digital and chip technology. However, nowadays, an Application Specific Integrated Circuit (ASIC) and the System-on-Chip (SoC) have made it much easier to build and control what are usually sophisticated and highly complex systems. ASIC and SoC comprise Very Large-Scale Integrated Circuit (VLSI) chips managed by a digital processor to convert application requirements into hardware operations. However, System-on-Chip (SoC) design faces several major challenges, with power consumption, design verification, and time-to-market pressure among the most pressing. Other key concerns include the growing complexity of designs and the need to effectively manage power delivery. The processor uses a hardware block that provides an interface to access targeted areas within the system using a dedicated bus protocol. Chip complexity, processing speed, power consumption, and chip size issues multiply when meeting these demands. Therefore, a single processor is insufficient for fast and more complex system applications. As a result, a new approach to multiprocessor access systems has emerged in chip manufacturing industries, solving these problems. The actual collective solutions are challenging for researchers and engineers. This research proposes and develops an efficient multiprocessor interface with a sophisticated arbitration scheme to address multiple processor access with higher speed, lower power and lower area to support efficient access by several processors in a system. Instead of typically using one type of arbitration for different processors, this research developed a dual-mode arbitration scheme for multiprocessor access to the system chip to support a broader range of system requirements. This dual-mode arbitration scheme manages access to shared resources such as the bus or memory in a system. Considering the design and fabrication complexity of multiprocessor interface hardware development, this study adopts a hierarchical development approach using the industry-standard Advanced Microcontroller BUS Architecture (AMBA) bus protocol, which is a highly accepted processing technique in the high-tech chip manufacturing industry. The research work is hardware modeled using the Verilog Hardware Description Language (HDL), and its performance is verified using Cadence and the ModelSim simulators; these tools offer comprehensive simulation and verification capabilities, reducing the need for costly physical prototypes and accelerating the design process. It is synthesized and implemented hardware using the Xilinx Synthesis Tool (XST), Electronic Design Automation (EDA) tool suite and Field Programmable Gate Array (FPGA) devices, a widely accepted alternative to synthesizing via custom fabrication in a foundry using Cadence synthesis tools and hardware implementations. Emulation, synthesis, hardware verification, and simulation clearly show that the design hardware achieves a 120% faster, 80% lower power, and 70% less area-efficient interface for multiprocessor access with an intelligent arbitration system. Therefore, for better throughput and faster data transfer, the results of this study can provide a smart solution of an integrated system for bulk data transfer that can be modeled as separate hardware for future work that can be combined or used as a separate intellectual property (IP), or can be done by a Macroblock in the hardware systems. These will encourage future researchers and entrepreneurs in the VLSI chip manufacturing industry.

Several recent studies have proved the ability of cement-based materials to capture carbon dioxide (CO2) through carbonation. Yet, the capture capacity may decline over multiple cycles, reflecting the poor regeneration performance inherent in other calcium-based sorbent materials. Partial replacement of nano-sized silica (SiO2) could potentially enhance both CO2 capture capacity and regeneration performance of cement. While previous research has extensively proven the significant improvement in cement properties with nano-silica, limited studies have examined its impact on CO2 capture and regeneration performance. Therefore, this study investigates how partial replacement of nano-silica in cement paste samples affects CO2 capture capacity and regeneration performance. Nano-silica was synthesized from rice husk ash (RHA) through the precipitation method, aiming to utilize agricultural waste. Before synthesizing, the RHA was acid-leached and thermally treated. Cement samples were partially replaced with nano-silica in various percentages (0.00% to 3.00%) and cured for 7, 14, and 28 days. Using the one-factor design from response surface method (RSM), the cement/nano-silica samples’ composition was determined. Characterization and analysis confirmed successful synthesis of high-purity, amorphous silica nanoparticles with diameters below 50 nm via the precipitation method. Nano-silica significantly improved the properties of hardened cement samples, with a notable 34.77% increase in compressive strength achieved with 3.00% nano-silica replacement compared to other samples across curing durations. XRD patterns indicated that nano-silica promoted hydration reactions, resulting in increased peak intensity of the C-S-H phase. Moreover, SEM-EDX analysis revealed the morphological characteristics of C-S-H phase throughout the observed morphology, along with a decrease in the Ca/Si ratio with increasing percentage of nano-silica replacement. The study findings suggest that inclusion of nano-silica significantly enhanced CO2 capture and regeneration performance of cement at room temperature conditions, with maximal improvement observed at 3.00% nano-silica partial replacement and 28 days of curing, displaying approximately 493.76% increment over the reference sample during a 150-minute experimental test. However, at 800℃ experimental temperatures, the presence of nano-silica did not effectively enhance CO2 capture capacity but rather led to its deterioration, potentially due to structural modification of the C-S-H phase during thermal cycles.

The study investigates the application of deep learning-based techniques for Autism Spectrum Disorder (ASD) diagnosis using facial image datasets, aiming to contribute innovative insights to scientific literature. The research focuses on developing a robust framework for ASD diagnosis utilizing Convolutional Neural Networks (CNNs) like Xception, MobileNetV2, and ResNet50V2. A key innovation lies in leveraging facial image features as potential biomarkers to distinguish between ASD and Normal Control (NC) children. This approach is supported by the capacity of deep learning to extract nuanced facial features imperceptible to human observation. The study employs transfer learning via both model-centric and data-centric approaches to analyze datasets, including the Kaggle and YTUIA datasets. Hyperparameter tuning on the Kaggle dataset with the Xception algorithm achieves optimal accuracy of 0.95, surpassing prior studies and establishing benchmark settings. Further employing the method on the YTUIA dataset enhances accuracy performance to 0.965 by ResNet50V2, encompassing a broader demographic range and enriching ASD research by YTUIA. Explainable AI reveals that Xception focuses on central facial regions for ASD diagnosis, while ResNet50V2 and MobileNetV2 rely on peripheral features. To validate findings, the study introduces the UIFID dataset, comprising 130 ASD and Normal control (NC) samples. Models trained on Kaggle and YTUIA exhibit validation accuracies ranging from 0.72 to 0.79 and 0.45 to 0.60, on UIFID respectively, emphasizing the challenge of cross-domain validation. Addressing these limitations, data-centric strategies incorporating pre-processing and augmentation achieve peak accuracies of 0.989 on Kaggle, though declining to 0.823 on UIFID, necessitating enhanced feature generalizability. Advanced deep learning methods, including active learning, federated learning, and ensemble learning, are employed to mitigate domain divergence. Active learning achieves accuracies of 0.80 and 0.773 on combined test set (accumulated from Kaggle and YTUIA) and UIFID datasets, respectively, demonstrating iterative model improvement. Federated learning achieves 0.95 accuracy on combined datasets and addresses ethical concerns, while ensemble learning surpasses all methods with 0.96 accuracy on combined datasets and 0.901 on unseen UIFID data using the Fifty Percent Priority Rule (FPPR) algorithm which enhances ensemble learning by prioritizing models based on validation accuracies.

The uncertainties associated with renewable Distributed Generation (DG) and distribution networks present significant challenges for researchers. DG systems, which integrate various renewable energy sources (RES), which consist of photovoltaics, wind turbines, hydrogen cells, and micro-turbines, are appealing in terms of economic and environmental advantages. Modular structural inverters are commonly used in DG, representing a new generation of systems that will soon integrate networks of RES. However, the modular structure and the integration of both AC and DC sources bring about stability challenges that must be pointed out. At the distribution substation, solar photovoltaic (PV) panels, microturbines, wind turbines, and hydrogen cells are connected to form a grid system known as DG. The integration of multiple RES, whether DC or AC, makes DG an attractive option due to its numerous economic and environmental benefits. This research proposes an inverter control mechanism that combines Quantitative Feedback Theory (QFT) and Maximum Power Point Tracker (MPPT) control to overcome uncertainties and ensure efficient controllability and stability in AC-DC hybrid DG systems. This research proposes a novel inverter control mechanism that can efficiently manage the output of power output and stability of an AC-DC hybrid MG system. In the simulation process, the simulation model was divided into several segmental modules interconnected with a controller. In this approach, the proposed inverter mechanism can connect both DC and AC modular RES and convert them to a fixed 400V DC. The Maximum Power Point Tracking (MPPT) control method is then applied to optimize the output power, while the Quantitative Feedback Theory (QFT) control mechanism adjusts power based on the demands from distributed generation (DG) or ESS. The excess power is stored in the energy storage systems if there is no immediate demand. The proposed model was simulated using MATLAB, and the results demonstrate the control strategy effect in maintaining grid stability and maximizing power output, outperforming the benchmark studies from RES. The innovation of this research lies in developing a comprehensive control mechanism capable of effectively managing the complexities of a modular structured RES microgrid system. The total harmonic distortion (THD) analysis is being conducted, keeping the THD percentage minimal. The research found the THD to be 3.80%, which is below the 5.00% standard, aligning with established norms. The proposed control strategy has proven its ability to point out the challenges due to uncertainties in renewable energy systems, making it a promising solution for integrating DG into the power grid.

In this research, open-channel bifurcations, defined as geometrical singularities of a hydrographic network, where the inflow separates into two downstream branches are investigated. Bifurcations are naturally present in river networks, mostly upstream from islands and in deltas, where the main flow does not necessarily continue in a straight direction but may form a dimensionally asymmetrical Y-shaped diverging channel. Sediment transport in bifurcations are increasingly becoming a major issue when extensive suspended as well as bed material conveyed by rivers regularly encroach into side or lateral (i.e. bifurcated) channels utilized for abstracting raw water thus affecting optimal intake for potable consumption. These divisions in flow are also observed in human-intervened conditions such as wastewater networks where the combined sewer overflows evacuate part of the exceeding sewer discharge capacity toward the surrounding environment. In addition, flooded streets reaching three-branch crossroads generate bifurcated channel flows. For most design or man-made cases, the main flow continues straight ahead, while the side branch, generally narrower, makes an angle close to 90° with the main channel, forming a junction known as T-shape bifurcation. Many researchers have completed studies on T-shape bifurcations to understand the behaviour of water flow and sediment transport. However, a complete understanding of the phenomenon, especially in relation to 2D or 3D secondary flows and vortices, is lacking up to this day. Moreover, the distribution of flow discharges in both main and lateral/branch channels requires further detailed investigation. As such, the objectives of this research on open-channel flow in 90° bifurcations are to determine experimental techniques of flow visualization, investigate methods to identify three-dimensional flow patterns, improve empirical discharge distribution relationships, and establish the effect of reduced side branch width on flow distribution. A total of 668 experimental runs have been conducted and completed in the open-channel flume system at Laboratoire de Mécanique des Fluides et d’Acoustique (LMFA) of Institut National des Sciences Appliquées (INSA) in Lyon, France. From the results, it is observed that flow visualization techniques were generally not effective nor adequately consistent to differentiate between 2D standard recirculation or 3D helical flow structures. Therefore, another approach based on analysis of fluid power gain or loss plotted against inflow discharge in the main channel is introduced as an alternative. As for assessing flow division through the bifurcation, the results are organized into sections to progressively address discharge distributions in the subcritical or free recirculation transcritical regimes. Consequently, a new predictive set of empirical equations for flow distribution is derived based on the known upstream discharge plus branch-width ratio and stage-discharge relationships in the downstream or lateral/branch channels. This discharge distribution formulation is accurate, as indicated by determination coefficients of at least 0.98 when compared to measurements, without even identifying beforehand whether the flow is in fully subcritical or free recirculation transcritical regime through the lateral/branch channel. Based on the successful accomplishment of stipulated objectives, recommendations for further work to include establishment of a flow-structure identification system, as well as extending the experiments to bifurcations with significantly narrower side branches. The angle of flow divergence could also be varied to other than 90°.

Major source of enzymes is microbes and during cultivation step, more than 85% of them resist cultivation and their genetic patrimony is loss. This work sought to bioprospect cellulose-degrading enzymes from microbiota of the palm oil mill effluent (POME). To get access to this great microbial diversity and to discover the biocatalysts behind, this research has adopted metagenomic approach which technically escapes the cultivation step and screens the microbial DNA for desired enzymatic activity. Metagenomics consists of the creation and screening of metagenomic DNA libraries. In vivo identification of cellulose-degrading enzymes was carried out with high-throughput screening and in silico identification of genes encoded the enzymes was performed with algorithm-based methods while the results validation was executed by recombinant enzymes expression, purification and characterisation. Culture-enrichment strategy based on natural selection principle was used in the early stage to enhance the screening hit rate. Metagenomic DNA was extracted from enriched and non-enriched sample to construct 109,824 fosmids (4.49 Gb) metagenomic DNA library. In this library, pCC1FOS fosmid and E. coli EPI300T1R are the vector and surrogate host of the cloning system, respectively. A high throughput functional metagenomic screen was developed and applied to search for cellulose-degrading enzymes within the library clones using 4-methylumbelliferyl-ß-D-glucopyranoside (MUGlc) and 4-methylumbelliferyl-ß-D-cellobioside (MUC) fluorogenic substrates. The screens were normalised using robust z-score and highest rated clones (100) were then selected. Their fosmids were isolated and sequenced with Hiseq (Illumina) of next-generation sequencing strategy. For quality control of the reads, SolexaQA and FastQC tools were used. Poor quality bases were removed with DynamicTrim algorithm, all bases with Qphred less than 20 were trimmed, and the LengthSort algorithm was used to remove sequences less than 50 bp. de Bruijn graph of de novo assembly algorithm has organised the reads on k-mers to build contigs, and Velvet optimiser has selected the optimum k-mers. These contigs were the input of SSPACE algorithm used to locate and orient contigs. Codon DNA sequences (CDS) were identified with PRODIGAL software. The genes identification was carried out following Blastp and SmartBLAST. Seventeen of bioprospected putative cellulose-degrading enzymes were cloned into pBAD-TOPO plasmid and expressed in TOP10 E.coli cloning. Enzymes were purified with HisTrap HP column with the aid of a FPLC system. Two putative glucanases and two putative ß-glucosidases were then biochemically characterised for optimal pH and temperature in the presence of substrates MUGlc and MUC, pNPG and pNPC and CMC as well. In NGS-data analysis step, 4900 contigs and 3540 scaffolds were constructed. 42,247 CDS were detected and 96 potential cellulose-degrading enzymes were identified which evinces the richness of POME metagenome on biocatalysts. The protein sequences of 15 cellulose-degrading enzymes are 100% similar to protein sequences available in protein databases while 40 enzymes show (80-99%) similarities, 24 enzymes (60-79%), 14 enzymes (40-59%) and 3 enzymes show less than 40% similarity, this reflects the qualification of functional metagenomics to bioprospect untapped enzymes. The potential types of enzymes are 19.20% glucanases, 31.32% glucosidases and 46.48% glucoside hydrolases with cellulose-degrading enzyme conserved domains. For the 17 expressed enzymes, three different glycoside hydrolase families, (enzyme 1, 2, 10 and 21) from GH3, (enzyme 6, 12 and 20) from GH5, (enzyme 3) from GH8 and other glycoside hydrolase families. Enzyme 3 is probably an example of untapped enzyme; it was active toward MUC and MUGlc. The optimum catalysis activity by enzyme 3 occurred at 50 °C and pH 4. For enzymes 4, 11 and 13, no enzymatic activity was detected due to low expression level. This research was very challenging but rewarding. It lays the foundation of diverse and untapped biocatalysts discovery. The bioprospected enzymes found in this metagenomic DNA library can be produced and optimised to be used in different industrial applications. In addition, the NGS-analysed-data can be usd to study the diversity of POME.

Microbial fuel cell (MFC) suffers from low energy gain yield (output/cost) and is unsuitable for most practical implementations. Microbial zinc/air cell employing freely suspended white rot fungal Phanerochaete chrysosporium fed with empty fruit bunch (EFB) has demonstrated promising prospects as a sustainable MFC. This work aimed to increase the energy gain yield of the system by minimizing external control features, implementing low-cost design, and increasing the energy output. To fulfil the power output for most low-power applications, multiple MFCs need to be connected in a stacking configuration. However, the variation in individual MFC’s electromotive force (e.m.f) due to living microbes activities induces parasitic currents in parallel configuration. This work introduced a novel open-parallel unit-cell configuration for MFC stacking. All unit cells were connected in parallel configuration but hydrodynamically connected i.e. they shared a common electrolyte. Using this configuration, the discharge capacity of the MFC stack was extended 3.4 times, the total power output was increased by 2.6 times compared to the common parallel configuration, and the parasitic current was effectively eliminated. The microbial zinc/air cell is an air-cathode MFC. The air cathode is the most expensive component in an air-cathode MFC and, in most cases, requires an air aeration system. Therefore, this work designed and fabricated a low-cost and easy-to-fabricate air cathode. It is low cost because it is non-catalytic and the fabrication did not require special processes, only mere mechanical press of the cell holders. Further, the air cathode components of carbon felt, carbon fibre sheet and nickel mesh, were designed for operating under submerged conditions and depending only on dissolved oxygen. Therefore, air aeration is not required. The proposed air cathode was capable of sustaining a discharge current of 1 mA for 42 days (1008 mAh) under submerged conditions thus supporting its viability. Aside from the low-cost design, the cylindrical air cathode configuration also offers the advantage of compact multipolar design. Since the microbial zinc/air cell was fed with lignin rich EFB as a substrate for Phanerochaete chrysosporium, this work assessed its efficacy as a lignin rich agrowaste degradation cell. The rates of lignin degradation were evaluated under self-generated current and externally supplied current. It was found that electric current stimulus enhances the lignin degradation. Externally supplied current induced higher lignin degradation. However, when the current supplied was 5 mA or higher, it disrupted the lignin degradation rates.

Food waste (FW) poses a significant global challenge due to the increasing population. Anaerobic digestion (AD) was a commonly used process to convert organic waste into biogas, primarily methane (CH4). However, CH4 production in AD was often low. To upgrade CH4 production, a hybrid microbial electrolysis cell (H-MEC) system combining AD was used. In this study, two types of fungal strains (TNAF-1 to TNFA-3 and TNBC-1 to TNBC-3) were isolated from animal feed and compost. The enzyme activities such as cellulase, and amylase of 300U/mL and 400U/mL, respectively were produced by the selected strains. Optimization using the face centered central composite design (FCCCD) under the response surface methodology (RSM) was conducted to increase the reducing sugar production to 162 mg/mL under the optimized conditions of pH 5, Total solids (TS) of 12.5%, and enzyme loading of 80 U/mL. The biogas production was optimized using one factor at a time (OFAT) method with the parameters of inoculum of 25%, pH 7, digestion times of 29 days, 500mL of hydrolysate food waste, and temperature at 30°C (±2), resulting in a biogas composition of 3% H2, 57% CH4, and 40% CO2. To address energy efficiency and sustainability, the H-MEC system was designed based on electromethanogenic microbes (EMMs) for enhanced biogas upgrading. EMM strains (TNFW-1 to TNFW-3 and NTAS-1 to NTAS-3) were isolated from AD using food waste and anaerobic sludge samples, respectively. The EMMs demonstrated the efficient CO2 conversion in a dual-chamber method. The strain, TNFW-2 produced supernatants with a 92% BioM (biomethane) production rate from the direct gas phase (CO2/H2). Optimal growth conditions were determined, yielding a 92% BioM yield with substrate dose of 100mL, inoculum dose of 10mL, flow rate of 5L/hour, H2/CO2 ratio of 50:50%, pH 7, applied potential of 900mV, and 36 hours of incubation. The SEM (spell out) images revealed irregular EMM structures with netted texture, and their activity was associated with a recorded potential value of 900mV. The strain, TNFW-2 was identified as the same species and the chemical composition of the extracellular EMMs was Methanobacterium formicicum of 98% sequence similarity. EMMs exhibited stability within a pH range of 4.5-8, with maximum CH4 production at a temperature of 28±2⁰C and pH 7 for at least 36 hours. The optimized H-MEC system demonstrated 92% CO2 conversion at an organic CO2 flow rate of 5 L/h and 36 hours of incubation time. The optimal H-MEC conditions for EMM-based biogas upgrading included two chambers with strainless steel (SS) with graphite (SS+GF) electrodes, an applied voltage of 900mV, pH 7, and an EMM dose of 10mL. Under these conditions, 92% CO2 removal in terms of CH4 production was achieved. Kinetic analysis revealed growth-associated BioM production, with an estimated specific growth rate (µ) of 0.207h-1 and maximum specific rate of product formation of approximately 0.239h-1. These findings highlight the potential of the new EMMs for non-toxic and biodegradable biogas upgrading, which may show a potential solution for future applications in the wastewater treatment plants.

The pressure in the base area of the enlarged duct and the backward-facing step of any moving projectile is usually less than the ambient pressure. It is essential to decrease the base drag by raising the base pressure. The present investigation is concerned with controlling base pressure using an annular cavity in the enlarged duct as a passive control mechanism. This investigative work will be valuable for space & defense and automobile engineering, as a stream is split at the blunt base. The computational fluid dynamics (CFD) investigation has been carried out in this research to examine the effect of expansion level, cavity size, and location on controlling base pressure. The variable quantities for the analysis are the nozzle pressure ratio (NPR), Mach number (M), cavity location (C), area ratio, and length-to-diameter (L/D) ratio of the duct. The theoretical framework establishes three conditions for the stream through a convergent-divergent nozzle: ideal expansion, under-expansion, and over-expansion. These conditions are intricately linked to the behavior of the boundary layer, reattachment length, and the strength of the primary vortex, all of which influence base pressure values. The study reveals nuanced insights by applying this framework to three specific area ratios (2.89, 4, and 5.29). For an area ratio of 2.89, cavity placement at 0.5D proves highly effective in regulating base pressure, showcasing the significance of Mach number, L/D ratio, NPR, and cavity location. The cavity's efficiency is attributed to its ability to counteract over-expansion conditions, impacting reattachment length and the strength of the primary vortex. For an area ratio of 4.0, the cavity demonstrates viability for base flow control at NPR = 4, with negligible impact at other NPR values due to over-expansion or highly under-expanded conditions. The study underscores the substantial influence of duct diameter, Mach number, L/D ratio, NPR, and cavity location on base pressure control, as evidenced in Taguchi's main effect plots. For an area ratio of 5.29, the cavity's effectiveness in manipulating base pressure is highlighted at NPR values of 4 and 6 for Mach numbers 1.2 and 1.4. The research underscores the intricate interplay between cavity placement, NPR values, and Mach numbers, revealing the complex relationship governing base pressure manipulation. In conclusion, this research provides valuable insights into the intricate flow dynamics within an enlarged duct. It emphasizes the pivotal role of cavity placement, NPR values, and Mach numbers in influencing base pressure, offering practical guidance for optimizing duct designs in engineering applications. The findings acknowledge the nuanced interactions of various parameters in controlling base pressure, contributing to a deeper understanding of this complex fluid dynamics phenomenon.

Pavement cracks pose a significant threat to road safety and infrastructure integrity, leading to potential hazards for vehicles and necessitating costly repairs. The traditional methods for detecting and characterizing these cracks are often manual, time-consuming, and subject to human error, making it challenging to maintain roads efficiently and effectively. This study advances pavement maintenance by applying deep learning and pixel-level segmentation to improve the detection and characterization (classification and sizing) of pavement cracks, crucial for road infrastructure safety and longevity. The research began with extensive data collection, using a GoPro Hero 8 mounted on a vehicle to gather images of roads in Kuala Lumpur and Selangor, Malaysia. Following detailed preprocessing, including image augmentation, blurring, and light intensity variations, a pavement crack dataset was prepared for analysis. The investigation started with a customized YOLOv7 model, achieving 0.9545 recall and 0.9523 precision on the RDD2022 dataset and 0.9158 recall and 0.93 precision on our custom dataset. When benchmarked against traditional methods, such as ConvNets and deep neural networks, the customized YOLOv7 model demonstrated better precision and recall performance. Subsequent work with the YOLOv8x-seg model resulted in improved precision and recall in crack detection, with performance metrics of 0.93 recall and 0.91 precision for alligator cracks, 0.95 recall and 0.84 precision for longitudinal cracks, and 0.806 recall and 0.89 precision for transverse cracks. In direct benchmarking, this model surpassed previous fusion-based deep convolutional methods in both precision and recall. The final phase involved creating an Advanced Hybrid Deep Learning Model incorporating Deep Gradient ResNet and a Modified Attention mechanism, enhancing crack detection and characterization. This model showed a promising precision of 0.84 for alligator cracks, 0.89 for longitudinal cracks, and 0.87 for transverse cracks, with recall rates of 0.96 for alligator cracks, 0.88 for longitudinal cracks, and 0.80 for transverse cracks. The developed model outperformed the benchmarked research utilizing CrackNet model in both precision and recall. Utilizing advanced segmentation and machine vision techniques, the study successfully demonstrated the capability to precisely size pavement cracks, which is vital for targeted maintenance. The research signifies progress in automated pavement crack analysis, contributing to safer and more durable road infrastructures.

The collagen family of proteins represents an abundant biopolymer that serves indispensable functions in maintaining the human body. Concerns over the halal/kosher status, sustainability, quality, and immunogenicity of animal-derived collagens have driven efforts to explore alternatives. Among these, recombinant bacterial collagen-like proteins emerged as prospective sustainable sources for various industries. This study investigates the scalability of a recombinant Rhodopseudomonas palustris collagen-like protein (RPCLP) production in E. coli. Before scale-up, the fermentation medium and process must be optimized to boost the growth and maximize RPCLP concentration. Hence, RPCLP was expressed in E. coli BL21(DE3) using the pColdII vector, and induction was accomplished using IPTG and temperature reduction. The optimal M9-casamino acid medium components were determined using a one-factor-at-a-time (OFAT) optimization. The screening of factors (using Taguchi OA) and, subsequently, optimization of selected factor levels (using FCCCD) to maximize RPCLP concentration were conducted in a 2 L bioreactor. Kinetic modeling was investigated by using the Monod and logistic growth models, and Luedeking-Piret models for the glucose consumption and RPCLP production. Nonlinear regression was used to determine kinetic parameters, and ANOVA analyses were conducted to evaluate the models. The fermentation was then upscaled into a 7.5 L stirred tank reactor (STR) via two scale-up criteria, namely constant oxygen mass transfer coefficient (constant kLa) and constant impeller tip speed. The soluble protein, obtained via ultrasonication, was characterized through trypsin digestion and SDS-PAGE, and purified using metal affinity chromatography. MALDI-TOF was also conducted to determine its molecular weight. The designs of the screening and optimization experiments were made by using Design-Expert software. The factors that significantly govern the growth of recombinant E. coli expressing R. palustris collagen-like proteins in the 2 L bioreactor were identified. The optimum conditions for the growth of the recombinant host cells and the concentration of recombinant collagen-like protein in the 2 L bioreactor were established. Optimum conditions yielded 2 g/L RPCLP, which is moderate compared to typical yields for recombinant collagen-like proteins. In the kinetics study, the Monod and logistic growth models fitted well, with significant R2 values (0.96 and 0.99, respectively). In contrast, the Luedeking-Piret models generated were not in accord with experimental data for glucose consumption and RPCLP production (R2 values ? 0.7). The kinetic parameters were determined, and the findings suggest that the production is partially growth-associated. The constant impeller tip speed criterion was effective as the maximum tip speed of 0.73 m/s prevents shear damage of the cells and the protein product. Contrarily, the constant kLa criterion was inadequate for scale-up, possibly due to heterogeneous environments caused by improper mixing and mass transfer during kLa measurement, leading to inaccurate estimates. Concentrations of approximately 2.1 to 2.5 g/L RPCLP were obtained in both scales using the constant impeller tip criterion. These concentrations were comparable to a previous study that obtained approximately 1 g/L recombinant Scl2 protein from a 2 L batch culture. A protein band that corresponds to the estimated size of the recombinant RPCLP from the literature was detected via SDS-PAGE. This study is the first to demonstrate successful medium optimization, kinetic modeling, and scale-up of recombinant RPCLP production to a 7.5 L bioreactor. The findings establish a protocol for an enhanced bioprocess to cater to the expanding market demands for sustainable, halal, or vegan collagens.

The reliability of smart textile garments for electrocardiogram (ECG) based biometric recognition presents a promising yet underexplored avenue in the development of wearable authentication technologies. While ECG signals offer unique physiological characteristics ideal for biometric verification, real-world deployment poses challenges related to signal variability, consistency, and classifier performance. This research directly addresses these issues by evaluating the reliability of two commercially available smart textile shirts of HeartIN Fit and Hexoskin ProShirt for ECG based biometric verification under practical and real-life conditions. Moreover, to thoroughly assess the approach robustness, the study was conducted through three structured experiments involving a total of 22 participants. The first experiment focused on distinctiveness, which evaluated how well the approach could identify individuals under varying physiological conditions such as walking, sitting, standing, and lying down. The second examined permanence, testing whether biometric performance remained stable across different time intervals, which is critical for long-term usability. The third experiment investigated system performance by benchmarking classification accuracy using a wide spectrum of 29 machine learning algorithms, including Quadratic Support Vector Machine (QSVM), Neural Networks (NN), and k-Nearest Neighbour (kNN). In order to ensure signal clarity and consistency, a low-pass Butterworth filter with a 30 Hz cut-off was applied to remove high-frequency noise while preserving key biometric features. Meanwhile, the HeartIN’s 512 Hz ECG signals were resampled to 256 Hz to align with Hexoskin’s sampling rate, enabling a normalised and fair comparative analysis. Subsequently, feature extraction stage captured the unique ECG morphology by detecting R-peaks and segmenting ten data points on each side, a process that successfully enhanced the reliability of biometric classification. Interestingly, the results demonstrated high reliability across all experiments, with the best performing classifier being QSVM, which achieved 98.81% accuracy with a false rejection rate (FRR) of 16.21% and a false acceptance rate (FAR) of 0.50% under different physiological conditions. When tested across time variability, it accomplished 99.20% accuracy with a FRR of 9.68% and a FAR of 0.27%. In addition, across the comparison with 29 classifiers, it attained 97.4% accuracy with a FRR of 2.60%, further confirming its robustness and reliability in diverse scenarios. Thus, these findings also provide strong evidence that ECG signals acquired through smart textile garments can deliver highly reliable biometric performance in dynamic and real-world environments. Therefore, through combining commercial wearable technology with broad machine learning evaluation, this study advances ECG biometrics and positions smart textiles as a secure, adaptable, and scalable platform for next-generation biometric verification.

Biometric plays a significant role in person verification. However, conventional biometric methods such as fingerprints and facial recognition are vulnerable to identity theft. Electrocardiogram (ECG) has emerged as a promising biometric modality due to its unique characteristics and inherent liveness detection criteria, making it difficult to forge. Preliminary studies have demonstrated the feasibility of ECG for biometrics recognition; however, challenges remain in ensuring its practicality, reliability, and usability in real-world scenarios. This thesis identifies three main research issues that are crucial for enhancing user acceptance of ECG based biometric recognition which are distinctiveness, collectability, and permanence. Existing studies predominantly focus on ECG biometrics in controlled environments. Therefore, this study explores biometric verification in two circumstances which are low-sampling ECG signals under different physiological conditions and compressed ECG signals without prior decompression considering four variables that are different activity settings, gender group, time variability and age categories. For low sampling data, Maximal Overlap Discrete Wavelet Transform (MODWT) is introduced as a novel signal pre-processing technique. The study managed to highlight the issue of distinctiveness where the outcome has demonstrated that different individuals under various physiological conditions possess each own QRS complex as unique identifiers with classification accuracy up to 99.7% by using Quadratic SVM. Secondly, to be able to perform biometric in a compressed state considering four different variables emphasises the following issues that are collectability and permanence. By using Symlet wavelet as the compression method, the classification accuracy by using Medium Gaussian SVM remains high. Different activity settings achieved a classification accuracy of up to 97.7%, demonstrating the robustness of ECG biometrics across various physical movements. Gender-based analysis shows an accuracy up to 98.6% where male participants show higher accuracy compared to female subjects. Time variability demonstrate high accuracy up to 98.8%. Additionally, age group analysis reveals that younger subjects achieve higher classification accuracy compared to older participants. The proposed approach demonstrates that even in a compressed state, ECG signals can still be effectively collected and measured and this addresses two important issues in biometric which are collectability and permanence. The study also discovered that, optimal compression of 50% can be achieved with high classification accuracy at level of decomposition of N=1. File size of the ECG signal also significantly decrease with each level of decomposition. Furthermore, each level of decomposition significantly reduces the ECG file size, making it more efficient for real-world applications without compromising verification performance. The robustness of the suggested technique is proven as the high classification results show that biometric recognition can be performed in compressed state. In conclusion, ECG based biometric is a promising approach to combat identity theft and complementing the current authentication system. By addressing the main issues of distinctiveness, collectability, and permanence, the proposed method suggests a robust and reliable mechanism for real life applications that offers a significant advancement in biometric recognition.

With the widespread use of Internet of Things (IoT) devices, the issue of designing efficient and secure routing protocols, including reasonable trust management, has attracted more attention in networking research. The proliferation of IoT-connected devices poses significant security challenges while enabling the evolution of IoT applications. It is challenging to implement security in such a constrained context given the IoT's integration of wireless sensor networks along with other distinctive features. When designing security mechanisms, notably routing protocols, lightweight approaches are preferred, as security is often considered pricy regarding processing power, energy, and memory. Secret key distribution schemes impose high computational costs and resource demands. The adoption of bio-inspired approaches contributes to discovering the optimal path for IoT routing by modeling the cognitive behavior of insect colonies to attain security cost-effectively for their inherent adaptable and scalable features. Bio-inspired strategies are shown to be robust, adaptive to environmental variations, and use less energy and computing power to offer optimal solutions for designing secure routing algorithms and autonomous distributed systems. In this thesis, a trust-aware secure bio-inspired WSN routing algorithm based on ant colony optimization (ACO) for IoT has been proposed and analyzed to find a secure and optimal routing path that is energy-efficient and not computationally burdensome while also aiming to provide trust in dynamic IoT environment. Efforts are made to enhance scalability, accommodate node mobility, and minimize initialization and processing delays to find the optimum forwarding path to overcome the existing issues for time-critical applications in IoT. In the proposed design methodology, an exclusive trust evaluation scheme through an improved version of ACO is introduced towards secure data transmission, optimizing the sensor’s residual energy and trust score. Based on how a node interacts with its network neighbors, a node's trustworthy behavior is predicted through the use of beta distribution. A trust-based threshold mechanism is used for node evaluation, and if the evaluated node’s computed direct trust value lacks credibility, the indirect trust value is determined. Neighbors with higher trust metric values are preferred for secure routing, but nodes with lower trust values are considered malicious. The artificial ants search for the most reliable path to route traffic through the next hop based on the higher energy factor and trust grades of the nodes. The fitness function includes the trust metric along with the current node energy and path cost to evaluate route performance. The energy parameter, the trust metric, and the average mobility of the nodes are added to the probability formula of the ACO algorithm. MATLAB has been used to evaluate the proposed routing algorithm’s performance. The assessment results demonstrate that it has minimized the average energy consumption by approximately 50% regardless of the increase in the number of nodes, making the algorithm lightweight and scalable, when compared to the standard bio-inspired algorithms and existing secure routing protocols. The simulation study was conducted to show the effectiveness of the proposed system in defending against Rank and Sybil attacks. The assessment results illustrate the membership grade, showing how the evaluation model captures the sharp changes in node behavior and is more responsive to malicious activity. It showed a decrease in end-to-end delay of around 40% and achieved faster convergence for the initialization delay, about 60% improvement to determine the best cost route. The proposed ACO approach addresses and rectifies the traditional issues of slow convergence and parameter sensitivity, which beset conventional ACOs. Based on the relevant data, the routing protocol provides a secure and globally optimal route and can efficiently balance energy consumption and security.

Chemical coagulants have been widely used in water treatment. However, their potential negative impacts on human health and the environment keep motivating researchers to develop commercially available natural coagulants. Microbial coagulants are widely recognized for their safety and water treatment capabilities, but their high production costs pose a significant challenge to their widespread use. Therefore, this study aims to develop a solid-state fermentation system to produce a myco-coagulant for water treatment. In an attempt to reduce the production, low-cost lignocellulosic biomasses were screened in this study among them coco peat was a suitable lignocellulolytic substrate for producing an efficient myco-coagulant with a high flocculating activity of 84.6 % without any optimization. Low nutritional condition is also an effective strategy to reduce the production cost; fermentation process parameters were optimized in this study in order to maximize myco-coagulant production, which was achieved after 5 days of incubation at 25º C and moisture content of 70 %, the highest flocculating activity of 93.06 % was obtained under the conditions: 3 % (w/w) malt extract, 2.5 % (w/w) glucose, and pH 7. The potential fungal strain used in this study was identified as Phanerochaete concrescens. Characterization of the produced myco-coagulant revealed that it was primarily a polysaccharide-like substance mainly composed of 189.6 (mg/L) protein, a total carbohydrate of 170.5 (mg/L), and a total sugar of 4.20 (g/L). Myco-coagulant displayed an irregular structure of a compact nature and the elemental composition of oxygen (40.9 %), carbon (30.5 %), and a low percentage of N (5.2 %), H (6 %), P (5.0 %), Ca (2.0 %), K (4.7 %), Na (3.2 %) and Mg (2.5 %). FTIR and GC/MS analysis further indicated the produced myco-coagulant containing several compounds from varying chemical groups, including phenol, 2,4-bis (1,1- dimethyl ethyl), hydroxyl −OH, amine NH2, and ester groups as the important functional groups. Zeta potential revealed that the produced myco-coagulant was an anionic group. The produced myco-coagulant was cation-independent and pH tolerant. Based on One Factor at Time (OFAT) obtained results, optimization of the most influential parameters was carried out by applying FCCCD under RSM to develop a second-order regression model for successful improvement in flocculation. The optimum flocculation was 96 % under 200 rpm rapid mixing speed during 2 min and 90 rpm slow mixing speed during 22 min; the appropriate myco-coagulant dosage and settling time for initial turbidity of 600 NTU were 12 ml and 30 mins, respectively. TSS, COD, NH3-N, and TN removal of the purified water were 90 %, 44 %, 56 %, and 81 %, respectively. Since commercialization and industrial applications of microbial coagulants are still lacking, this study was the first attempt to scale up the production by applying different models; myco-coagulant successfully achieved a maximum flocculating rate of 92 % in foil tray under moisture content, pH value, and thickness of the substrate of 70 %, 7 and 3 cm, respectively. However, the large-scale production in an agitated bioreactor drum (ABD) was an attempt, and results show poor flocculability due to poor fungal growth in the presence of agitation rate at the vessel. The study reveals the potential of a safe, environmentally friendly myco-coagulant for sustainable water treatment, aligning with global environmental awareness and green technology for scaling-up purposes through solid-state fermentation.