Experimental Outcomes Research Articles

Robust detection and refinement of saliency identification

Salient object detection is an increasingly popular topic in the computer vision field, particularly for images with complex backgrounds and diverse object parts. Background information is an essential factor in detecting salient objects. This paper suggests a robust and effective methodology for salient object detection. This method involves two main stages. The first stage is to produce a saliency detection map based on the dense and sparse reconstruction of image regions using a refined background dictionary. The refined background dictionary uses a boundary conductivity measurement to exclude salient object regions near the image's boundary from a background dictionary. In the second stage, the CascadePSP network is integrated to refine and correct the local boundaries of the saliency mask to highlight saliency objects more uniformly. Using six evaluation indexes, experimental outcomes conducted on three datasets show that the proposed approach performs effectively compared to the state-of-the-art methods in salient object detection, particularly in identifying the challenging salient objects located near the image's boundary. These results demonstrate the potential of the proposed framework for various computer vision applications.

A pyramidal residual attention model of short‐term wind power forecasting for wind farm safety

AbstractWind power fluctuation significantly impacts the safe and stable operation of the wind farm power grid. As the installed capacity of grid‐connected wind power expands to a certain threshold, these fluctuations can detrimentally affect the wind farm's operations. Consequently, wind power prediction emerges as a critical technology for ensuring safe, stable and efficient wind power generation. To optimize power grid dispatching and enhance wind farm operation and maintenance, precise wind power prediction is essential. In this context, we introduce a joint deep learning model that integrates a compact pyramid structure with a residual attention encoder, aiming to bolster wind farm operational safety and reliability. The model employs a compact pyramid architecture to extract multi‐time scale features from the input sequence, facilitating effective information exchange across different scales and enhancing the capture of long‐term sequence dependencies. To mitigate vanishing gradients, the residual transformer encoder is applied, augmenting the original attention mechanism with a global dot product attention pathway. This approach improves the gradient descent process, making it more accessible without introducing additional hyperparameters. The model's efficacy is validated using a dataset from an actual wind farm in China. Experimental outcomes reveal a notable enhancement in wind power prediction accuracy, thereby contributing to the operational safety of wind farms.

OPTIMIZATION OF THE PEROVSKITE SOLAR CELL STRUCTURE

An experimental perovskite solar cell (PSC) with the structure Au/Spiro-MeOTAD/CH3NH3PbI3/PEDOT:PSS/ITO was fabricated. The measurements of main photovoltaic characteristics were provided. The current-voltage dependences (I-V curves) were measured conducted in the voltage range from -1V to 1V. During the measurements, the corresponding values were calculated of the short-circuit current density(Jsc) and open-circuit voltage(Uoc) were obtained as 1.23 mA/cm² and 0.19 V, respectively. Subsequently, an analytical model corresponding to this structure was formulated. For modeling the parameters of the perovskite solar cell, the Comsol Multiphysics environment was used, this environment is based on the finite element method. The relevant computations were provided to obtain the corresponding values of the short-circuit current density and open-circuit voltage as 3.29 mA/cm² and 0.2 V, respectively, with the maximum theoretically calculated power of this structure being 0.11 W. The experimental outcomes were juxtaposed with the predictions of the analytical computations, and the modeling results were empirically validated. An analytically accomplished model of the same structure was built by adding an electron transport layer (ETL). An organic material BCP (Bathocuproine) was used as an supplementary ETL layer. During the optimization of the PSC, the main datums were mathematically counted. Such values as the short-circuit current density of 10.17 mA/cm², open-circuit voltage of 1.2 V, and the maximum power value of Au/BCP/Spiro-MeOTAD/CH3NH3PbI3/PEDOT:PSS/ITO structure, which is 3.21 W were rated. A comparison of the volt-ampere characteristics of perovskite cells in dark and light modes was conducted for primary and optimized structures. The main parameters, obtained during the modeling of the experimental sample and subsequent model optimization, were compared. Specifically one of the key parameters of solar cell heterostructures the fill factor was evaluated and found to have increased from 16.52% to 25.00%, respectively. The light-sensitive behavior of the perovskite cell were visibly enhanced.

Efficient Geometric Correction Workflow for Airborne Hyperspectral Images through DEM-Driven Correction Techniques

Abstract. Geometric correction, a pivotal step in the preprocessing of airborne remote sensing imagery, is critical for ensuring the accuracy of subsequent quantitative analyses. Achieving precise and efficient geometric correction for airborne hyperspectral data remains a significant challenge in the field. This study presents a new method for system-level and fine-scale geometric correction of uncontrolled airborne images utilizing DEM data, which integrates forward and inverse transformation algorithms. Furthermore, an optimized workflow is proposed to facilitate the processing of large-scale hyperspectral datasets. The effectiveness of the proposed method is demonstrated through an application analysis using airborne HyMap imagery, with experimental outcomes indicating high application accuracy and enhanced processing efficiency.

Strength classes of brazilian hardwoods for structural design

For the elaboration of projects on timber structures, the Brazilian standard (ABNT, in Portuguese Associação Brasileira de Normas Técnicas) 7190 (ABNT 1997) ensures the correct application of physical-mechanical properties according to strength classes of lignocellulosic materials. This procedure eliminates the need for botanical identification of woods, since these strength classes support the efficient utilization of a wide range of woody varieties available in Brazil. Due to high mechanical resistances, the hardwoods are usually applied for structural projects of timber construction. This Brazilian standard document prescribes four strength classes (C20, C30, C40 and C60) for these woods, which are determined by characteristic value from the compressive strength parallel to the grain (fc0,k). But, these classes were obtained from experimental outcomes using a few wood varieties. The reorganization of strength classes should result in the best use of mechanical potentials of woods since the updating of these categories leads to the optimization of structural projects for timber construction in order to reposition this biomaterial at even more competitive levels. In this context, the present study aims to verify the current strength classes of hardwoods and if they lead to a good allocation of the fc0,k characteristic values. Otherwise, new strength classes can be determined for better allocations for efficient structural uses. As a result, 56 hardwoods were considered using 672 experimental determinations. Statistically, the current categories can lead to 12,5 % of incorrectly allocated values. The inclusion of C50 and C70 classes allow greater representation for these categories, in order to optimize the use of the hardwoods given the new strength classes. These findings ought to support future revisions of this Brazilian standard document.

Effects of handling and surgical implant on juvenile wild trout: Physiological and behavioral impact.

Fish involved in telemetry studies are usually handled, anaesthetized, and subjected to internal tag implantation, all of which have the potential to disrupt the fish's physiology, migratory patterns, food-seeking behavior, growth, and survival. As fish tagging is a widespread standard method in aquaculture, it is crucial to gain a better understanding of the short- and long-term effects on various aspects of fish welfare. The experimental outcomes of each phase of the surgical process and how it affects the fish's ability to recover during the post-tagging period, particularly in small salmonids, are poorly understood. Thus, it is essential to measure the extent of these adverse effects to precisely extrapolate differences in surgical intervention for tagged fish compared to their unmanipulated counterparts. In our study, we explored the post-tagging impacts on wild-captured brown trout (Salmo trutta) (1+) juveniles. We examined the time effect on fish recovery after each phase of the surgical procedure, which included anaesthetized fish, followed by surgery with/without internal tagging, within the first 2 and 25 h post-surgery. Common blood parameters (glucose levels and haematocrit) and endpoints related to swimming behavior, including fish movement, direction, and social distancing aspects, were studied. The findings of the study indicated no significant changes in glucose and haematocrit levels over the observational period. Fish subjected to anesthesia and tagging procedures exhibited a significant decrease in swimming activity, recovering to baseline levels 1 h after anesthesia and 24 h after surgical (tagging) procedures in recirculating aquaculture systems. Tagged fish showed less social proximity than non-tagged fish in the first 2 h post-surgery. A significant effect size was observed between nonsurgical (anesthesia) and surgical (non-tagged/tagged) groups, highlighting a more pronounced impact associated with surgical tag interventions. Our data indicate that the magnitude of behavioral response was significantly influenced by tag weight (~1.4% body/tag weight ratio in water) among manipulated fish groups. Consequently, our study reveals that wild-captured S. trutta juveniles, tagged with commercially available low-weight acoustic tags (V6, 69 kHz), experienced negative effects on swimming performance. In surgical studies, acknowledging potential influences is crucial for accurately inferring fish physiological and behavioral status. Emphasizing fish recovery potential in both short- and long-term periods is essential for quantifying tagging effects. Future research should prioritize exploring alternative tagging technologies and refining methodologies, with a particular focus on assessing telemetry's impact on socioeconomically relevant small salmonids.

Modeling carbonation depth of recycled aggregate concrete containing chlorinated salts

The durability of recycled concrete is susceptible to degradation from both chloride salt attack and carbonation when exposed in service environment. The depth of carbonation in recycled concrete subject to chloride salt attack was studied here. This investigation assessed the effects of mineral admixture, replacement rate of recycled coarse aggregate, and quality of recycled coarse aggregate on the carbonation depth. The carbonation depth Recycled concrete was studied through accelerated simulated carbonation in the laboratory for 7d, 14d and 28d. Based on the results, the correlation analysis was conducted to investigate the relationship between depth of carbonation, mineral admixture, recycled coarse aggregate replacement rate, recycled coarse aggregate quality and the presence of chloride ions. The objective of this study was mainly to investigate the effect of internal chloride ions on the depth of carbonation of recycled concrete subjected to chloride salt attack. The study referred to previous academic research on carbonation in recycled concrete that had not been affected by chloride salts, and conducted a comparative analysis to identify the carbonation depth pattern in recycled concrete influenced by chloride salt exposure. The study discovered a negative correlation between the replacement rate of mineral admixture and recycled coarse aggregate and the carbonation resistance of concrete. Moreover, enhancing the quality of recycled coarse aggregate positively impacted the carbonation resistance of concrete. An improved model was developed that considers the effect of internal chloride salts on the carbonation depth of recycled concrete, in reference to the empirical prediction model presently available. The model demonstrated a high level of agreement with experimental outcomes, displaying an error range of between −4.32% and 8.27%. It enables an in-depth comprehension of the carbonation behavior of recycled aggregate concrete in an environment exposed to chloride salt attack. Based on the model, the service life of recycled concrete structures can be more accurately forecasted, which can provide a theoretical direction for enhancing the durability of engineering structure.

Application of modified kaolinite in improving thermal stability of polydimethylsiloxane

AbstractThis research investigated the chemical structure and morphology of the hydrogenated tallow amine (HTA)‐intercalated kaolinite by XRD, FT‐IR, and TEM measurements. The experimental outcomes show that intercalation of kaolinite by HTA, with an intercalation rate of 63.2%, expands the interlayer distance with a subsequent dominant basal XRD peak at 2θ ≈ 7.9°. Furthermore, a composite was created from the HTA‐intercalated kaolinite and polydimethylsiloxane (PDMS) through a favorable interaction between the end silanol groups of PDMS and the modified kaolinite surface. Thermogravimetry‐differential scanning calorimetry results indicate that introducing the HTA‐intercalated kaolinite into the PDMS matrix increases the thermal stability of the composite from 178 to 343 °C. However, the thermal stability of the composite is lower than that of other published materials. The synthesis of kaolinite/PDMS composite with enhanced structural and thermal stability outcomes has been provided via a simple and effective solution method that should have further applications.

PEPPER AND CORN LEAVES CLASSIFICATION AND SEVERITY IDENTIFICATION USING HYBRID OPTIMIZATION BASED U-NET MODEL

Abstract An agricultural product plays a major role in the economical growth of developing countries. Agricultural products like pepper and corn are the essential crops with respect to human health food security. But, these two crops are prone to different diseases such as gray leaf spot, common rust and fruit rot which affects the productivity of crops. Further, the severity identification is also a challenging one. To address these limitations, this work presents different approaches for identifying the crop lesions and predicting the severity and thereby increasing the productivity of crops. The development of the proposed model includes steps such as dataset collection, noise removal, segmentation, feature extraction, classification and severity prediction. Initially, the crop images are pre-processed by the median filter and the pre-processed images are processed are segmented, extracted and classified by the optimized U-Net model. Moreover, hybrid optimizer which is the integration of GJA (Golden jackal algorithm) and RDA (Red deer algorithm) are utilized for precise segmentation and classification. Finally, the severity prediction is computed for the diseased leaves by the measuring the size of image pixels. The experimentation is carried out on the PlantVillage dataset; the accuracy and precision values achieved are 99.2% and 99.1%. Thus, the experimental outcomes show the effectiveness and stability of the model.

Synthesis, characterization and LDA+U calculations of zinc oxide nanoparticles

Zinc oxide (ZnO) was prepared by the sol-gel method as thin films deposited using spray pyrolysis. The characterization of the synthesized ZnO has been carried out using x-ray diffraction (XRD), Fourier-transform infrared (FTIR), scanning electron microscopy (SEM), and UV-Visible absorption spectroscopy. The experimental results revealed that the prepared ZnO has a wurtzite hexagonal structure with an average crystallite size of about 26.7 nm. High purity and flake-like structures were achieved as SEM and energy dispersive indicated. FTIR confirmed the prepared ZnO’s high purity as the Zn-O stretching peak was very intense. The optical parameters were comprehensively investigated, including absorption, reflection, extinction coefficient, refractive index, and optical energy gap. The wurtzite hexagonal structure of ZnO was optimized using the local-density approximation with the Hubbard correction method (LDA+U). Our experimental result for the energy gap is 3.28 (eV), which is in excellent agreement with the first principle calculations. We utilize the results from the LDA+U calculation along with our experimental outcomes to calculate the thickness of the thin film in UV-Vis spectroscopy.

Data-Driven Face Mask Detection and Alert System Utilizing Haar Cascade Classifier

The COVID-19 pandemic has brought to light the pivotal role of face masks in curbing the transmission of contagious illnesses. As societies strive to uphold mask mandates in public areas, there has been a surge of interest in automated face mask detection systems. This study presents a novel data-centric methodology for crafting a reliable face mask detection and notification system leveraging the Haar cascade classifier. Our approach hinges on a meticulously curated dataset comprising images of individuals both with and without masks. Through the adept utilization of feature extraction techniques and machine learning algorithms, our system learns to discern masked faces from their unmasked counterparts in real-time scenarios. The experimental outcomes of our study attest to the efficacy of our proposed system, showcasing commendable accuracy in detecting face masks. This substantiates its viability for deployment across diverse public settings, bolstering the arsenal of public health safety measures. The seamless integration of our system stands poised to fortify efforts aimed at safeguarding community well-being amidst the persistent threat of infectious diseases. Key words: Data, Mask , Deep Learning, Hybrid.

The relationship between austenite stabiliser element and post-weld heat treatment on the corrosion behaviour of the welded duplex stainless steel made by gas tungsten arc welding

The corrosion behaviour and the influence of nickel and post-weld heat treatment (PWHT) on the susceptibility to corrosion in welded duplex stainless steels were investigated. After PWHT, the amount of nickel for the weld region increases resulting in the generation of austenite and consequently, helping to obtain a favourable ferrite-to-austenite ratio. Comparing the experimental outcomes, it is noticed that the pits initiate and propagate in HAZ indicating this area is more susceptible to corrosion resistance compared to the fusion zone. The pitting resistance equivalent number alone is not sufficient reason to study the complex corrosion phenomenon, and the influence of other parameters namely microstructural evolution on the breakdown of passivity by pitting and the existence of intermetallic phases should be determined together.

3D-AMM: a 3D artificial moment method for path planning of manipulator in multiple obstacles scenario

Purpose In response to the issue of traditional algorithms often falling into local minima or failing to find feasible solutions in manipulator path planning. The purpose of this paper is to propose a 3D artificial moment method (3D-AMM) for obstacle avoidance for the robotic arm's end-effector. Design/methodology/approach A new method for constructing temporary attractive points in 3D has been introduced using the vector triple product approach, which generates the attractive moments that attract the end-effector to move toward it. Second, distance weight factorization and spatial projection methods are introduced to improve the solution of repulsive moments in multiobstacle scenarios. Third, a novel motion vector-solving mechanism is proposed to provide nonzero velocity for the end-effector to solve the problem of limiting the solution of the motion vector to a fixed coordinate plane due to dimensionality constraints. Findings A comparative analysis was conducted between the proposed algorithm and the existing methods, the improved artificial potential field method and the rapidly-random tree method under identical simulation conditions. The results indicate that the 3D-AMM method successfully plans paths with smoother trajectories and reduces the path length by 20.03% to 36.9%. Additionally, the experimental comparison outcomes affirm the feasibility and effectiveness of this method for obstacle avoidance in industrial scenarios. Originality/value This paper proposes a 3D-AMM algorithm for manipulator path planning in Cartesian space with multiple obstacles. This method effectively solves the problem of the artificial potential field method easily falling into local minimum points and the low path planning success rate of the rapidly-exploring random tree method.

Research on Multi-Objective Process Parameter Optimization Method in Hard Turning Based on an Improved NSGA-II Algorithm

To address the issue of local optima encountered during the multi-objective optimization process with the Non-dominated Sorting Genetic Algorithm II (NSGA-II) algorithm, this paper introduces an enhanced version of the NSGA-II. This improved NSGA-II incorporates polynomial and simulated binary crossover operators into the genetic algorithm’s crossover phase to refine its performance. For evaluation purposes, the classic ZDT benchmark functions are employed. The findings reveal that the enhanced NSGA-II algorithm achieves higher convergence accuracy and surpasses the performance of the original NSGA-II algorithm. When applied to the machining of the high-hardness material 20MnCrTi, four algorithms were utilized: the improved NSGA-II, the conventional NSGA-II, NSGA-III, and MOEA/D. The experimental outcomes show that the improved NSGA-II algorithm delivers a more optimal combination of process parameters, effectively enhancing the workpiece’s surface roughness and material removal rate. This leads to a significant improvement in the machining quality of the workpiece surface, demonstrating the superiority of the improved algorithm in optimizing machining processes.

Microwave-Solvothermal Synthesis of Mesoporous CeO2/CNCs Nanocomposite for Enhanced Room Temperature NO2 Detection

Nitrogen dioxide (NO2) gas sensors are pivotal in upholding environmental integrity and human health, necessitating heightened sensitivity and exceptional selectivity. Despite the prevalent use of metal oxide semiconductors (MOSs) for NO2 detection, extant solutions exhibit shortcomings in meeting practical application criteria, specifically in response, selectivity, and operational temperatures. Here, we successfully employed a facile microwave-solvothermal method to synthesize a mesoporous CeO2/CNCs nanocomposite. This methodology entails the rapid and comprehensive dispersion of CeO2 nanoparticles onto helical carbon nanocoils (CNCs), resulting in augmented electronic conductivity and an abundance of active sites within the composite. Consequently, the gas-sensing sensitivity of the nanocomposite at room temperature experienced a notable enhancement. Moreover, the presence of cerium oxide and the conversion of Ce3+ and Ce4+ ions facilitated the generation of oxygen vacancies in the composites, thereby further amplifying the sensing performance. Experimental outcomes demonstrate that the nanocomposite exhibited an approximate 9-fold increase in response to 50 ppm NO2 in comparison to pure CNCs at room temperature. Additionally, the CeO2/CNCs sensor displayed remarkable selectivity towards NO2 when exposed to gases such as NH3, CO, SO2, CO2, and C2H5OH. This straightforward microwave-solvothermal method presents an appealing strategy for the research and development of intelligent sensors based on CNCs nanomaterials.

A multimodal transfer learning framework for the classification of disaster-related social media images

Rapidly classifying disaster-related social media (SM) images during a catastrophe event is critical for enhancing disaster response efforts. However, the biggest challenge lies in acquiring labeled data for an ongoing (target) disaster to train supervised learning-based models, given that the labeling process is both time-consuming and costly. In this study, we address this challenge by proposing a new multimodal transfer learning framework for the real-time classification of SM images of the target disaster. The proposed framework is based on Contrastive Language-Image Pretraining (CLIP) model, jointly pretrained on a dataset of image-text pairs via contrastive learning. We propose two distinct methods to design our classification framework (1) Zero-Shot CLIP: it learns visual representations from images paired with natural language descriptions of classes. By utilizing the vision and language capabilities of CLIP, we extract meaningful features from unlabeled target disaster images and map them to semantically related textual class descriptions, enabling image classification without training on disaster-specific data. (2) Linear-Probe CLIP: it further enhances the performance and involves training a linear classifier on top of the pretrained CLIP model’s features, specifically tailored to the disaster image classification task. By optimizing the linear-probe classifier, we improve the model’s ability to discriminate between different classes and achieve higher performance without the need for labeled data of the target disaster. Both methods are evaluated on a benchmark X (formerly Twitter) dataset comprising images of seven real-world disaster events. The experimental outcomes showcase the efficacy of the proposed methods, with Linear-Probe CLIP achieving a remarkable 7% improvement in average F1-score relative to the state-of-the-art methods.

Theoretical analysis for the mechanical properties of the knitted structures

Abstract Knitting is a technology that has a thousand-year-old history, and can be normally seen in our daily lives. The knitted structure is constructed by the interwoven yarns that constrained by themselves, exhibiting extreme stretchability. The mechanical properties of the knit fabric also enable their integration with the flexible electronic devices. Nonetheless, it is yet problematic to expose the mechanical behaviors of knitting intrinsically. This paper investigates the mechanical characteristics of knitted structures subjected to uniaxial stretching. The analysis includes a structural assessment of the unit cell, with a focus on half of the cell accounting for symmetry. Mechanical analysis for three distinct scenarios (without elongation and friction, with elongation and no friction, with elongation and friction) is also presented. The stress-strain curve of the knitted structure and the correlation between stiffness and geometric parameters are illustrated. Additionally, simulations are carried out based on finite element analysis, yielding consistent results with the theoretical calculations. Subsequently, a uniaxial stretching experiment is conducted, and the experimental outcomes also verifies the theoretical analysis. Our analysis successfully explains the mechanical behavior of knitted structures, and also provides a reference for studying knitted fabrics with other topologies.

Chaotic-Based Mountain Gazelle Optimizer for Solving Optimization Problems

The Mountain Gazelle Optimizer (MGO) algorithm has become one of the most prominent swarm-inspired meta-heuristic algorithms because of its outstanding rapid convergence and excellent accuracy. However, the MGO still faces premature convergence, making it challenging to leave the local optima if early-best solutions neglect the relevant search domain. Therefore, in this study, a newly developed Chaotic-based Mountain Gazelle Optimizer (CMGO) is proposed with numerous chaotic maps to overcome the above-mentioned flaws. Moreover, the ten distinct chaotic maps were simultaneously incorporated into MGO to determine the optimal values and enhance the exploitation of the most promising solutions. The performance of CMGO has been evaluated using CEC2005 and CEC2019 benchmark functions, along with four engineering problems. Statistical tests like the t-test and Wilcoxon rank-sum test provide further evidence that the proposed CMGO outperforms the existing eminent algorithms. Hence, the experimental outcomes demonstrate that the CMGO produces successful and auspicious results.

Advancing Rock-Socketed Pile Design with a Unified Interface Shear Strength Framework for Soft Rocks

AbstractThe shaft resistance of rock-socketed piles (RSPs) is primarily influenced by the interactions between the pile, rock and any soft interface materials. This study presents a fundamental experimental and numerical investigation to predict the shaft resistance of model RSPs in soft rocks through a comprehensive shear strength framework incorporating the major variables such as roughness and smear fabric. By calibrating the Discrete Element Method (DEM) results with the experimental outcomes, this study evaluates the load-resistance attributes of RSPs in three different soft rocks for a wide range of roughness and smear configurations. The interface roughness effect of the RSPs in terms of the friction coefficient was correlated with the ultimate shaft resistances, uniaxial compressive strength and Young’s modulus of the rock. The study then incorporated the effect of smear fabric (placement, thickness and area proportion) in the shear strength framework of clean shafts. Comprehensive review of the DEM results revealed that the socket roughness effect diminishes at a critical roughness factor $$({\text{RF}})$$ ( RF ) of 0.4, beyond which the smear predominantly influences the shaft capacity. Following this, the effects of roughness and smear fabric were incorporated into a single equation representing the interface shear strength of RSPs, where the existence of smear results in a maximum reduction of up to 75% of the ultimate shaft resistance. The distinctive feature of this unified interface shear strength framework lies in its integration of the new smear fabric parameter and the linkage to the mechanical properties of soft rocks, which is the limitation of the earlier studies. It thereby sets a strong base for future studies aimed at advancing the pile-rock interface models.

A study on the effect of the geometric properties and surface defects on silicon chip flexibility for wearable electronics

Purpose The purpose of this paper was to investigate how variations in the geometry of silicon chips and the presence of surface defects affect their static bending properties. By comparing the bending radius and strength across differently sized and treated chips, the study sought to understand the underlying mechanics that contribute to the flexibility of silicon-based electronic devices. This understanding is crucial for the development of advanced, robust and adaptable electronic systems that can withstand the rigors of manufacturing and everyday use. Design/methodology/approach This study explores the impact of silicon chip geometry and surface defects on flexibility through a multifaceted experimental approach. The methodology included preparing silicon chips of three distinct dimensions and subjecting them to thinning processes to achieve a uniform thickness verified via scanning electron microscopy (SEM). Finite element method (FEM) simulations and a series of four-point bending tests were used to analyze the bending flexibility theoretically and experimentally. The approach was comprehensive, examining both the intrinsic geometric factors and the extrinsic influence of surface defects induced by manufacturing processes. Findings The findings revealed a significant deviation between the theoretical predictions from FEM simulations and the experimental outcomes from the four-point bending tests. Rectangular-shaped chips demonstrated superior flexibility, with smaller dimensions leading to an increased bending strength. Surface defects, identified as critical factors affecting flexibility, were analyzed through SEM and atomic force microscopy, showing that etching processes could reduce defect density and enhance flexibility. Notably, the study concluded that surface defects have a more pronounced impact on silicon chip flexibility than geometric factors, challenging initial assumptions and highlighting the need for defect minimization in chip manufacturing. Originality/value This research contributes valuable insights into the design and fabrication of flexible electronic devices, emphasizing the significant role of surface defects over geometric considerations in determining silicon chip flexibility. The originality of the work lies in its holistic approach to dissecting the factors influencing silicon chip flexibility, combining theoretical simulations with practical bending tests and surface defect analysis. The findings underscore the importance of optimizing manufacturing processes to reduce surface defects, thereby paving the way for the creation of more durable and flexible electronic devices for future technologies.